MARKETING METRICS
THIRD EDITION
This page intentionally left blank
MARKETING
METRICS
THIRD EDITION
THE MANAGER’S GUIDE TO
MEASURING MARKETING
PERFORMANCE
Neil T. Bendle
Paul W. Farris
Phillip E. Pfeifer
David J. Reibstein
Publisher: Paul Boger
Editor-in-Chief: Amy Neidlinger
Executive Editor: Jeanne Glasser Levine
Editorial Assistant: Kristen Watterson
Cover Designer: Chuti Prasertsith
Managing Editor: Kristy Hart
Project Editor: Andy Beaster
Proofreader: Debbie Williams
Indexer: Tim Wright
Compositor: Nonie Ratcliff
Manufacturing Buyer: Dan Uhrig
© 2016 by Pearson Education, Inc.
Publishing as Pearson
Upper Saddle River, New Jersey 07458
For information about buying this title in bulk quantities, or for special sales opportunities (which may
include electronic versions; custom cover designs; and content particular to your business, training goals,
marketing focus, or branding interests), please contact our corporate sales department at corpsales@
pearsoned.com or (800) 382-3419.
For government sales inquiries, please contact governmentsales@pearsoned.com.
For questions about sales outside the U.S., please contact international@pearsoned.com.
Company and product names mentioned herein are the trademarks or registered trademarks of their
respective owners.
All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in writing from the publisher.
Printed in the United States of America
First Printing September 2015
ISBN-10: 0-13-408596-5
ISBN-13: 978-0-13-408596-8
Pearson Education LTD.
Pearson Education Australia PTY, Limited
Pearson Education Singapore, Pte. Ltd.
Pearson Education Asia, Ltd.
Pearson Education Canada, Ltd.
Pearson Educación de Mexico, S.A. de C.V.
Pearson Education—Japan
Pearson Education Malaysia, Pte. Ltd.
Library of Congress Control Number: 2015943269
We dedicate this book to our students, colleagues,
and consulting clients who convinced us that
a book like this would fill a real need.
This page intentionally left blank
CONTENTS
Acknowledgments ix
About the Authors xi
Foreword xiii
Foreword to Third Edition xv
1
INTRODUCTION
1
2
SHARE OF HEARTS, MINDS, AND MARKETS
17
3
MARGINS AND PROFITS
67
4
PRODUCT AND PORTFOLIO MANAGEMENT
111
5
CUSTOMER PROFITABILITY
157
6
SALES FORCE AND CHANNEL MANAGEMENT
185
vii
7
PRICING STRATEGY
225
8
PROMOTION
271
9
ADVERTISING METRICS
295
10
ONLINE, EMAIL, AND MOBILE METRICS
325
11
MARKETING AND FINANCE
363
12
THE MARKETING METRICS X-RAY AND TESTING
13
SYSTEM OF METRICS
401
Bibliography 417
Endnotes 421
Index 429
viii
MARKETING METRICS
383
ACKNOWLEDGMENTS
We hope this book will be a step, however modest, toward clarifying the language,
construction, and meaning of many of our important marketing metrics. If we have
succeeded in making such a step, we owe thanks to a number of people.
Jerry Wind reviewed our initial concept and encouraged us to set our sights higher.
Rob Northrop, Simon Bendle, and Vince Choe read early drafts and gave valuable
feedback on the most important chapters. Eric Larson, Jordan Mitchell, Tom Disantis,
and Francisco Simon helped develop material for important sections and provided
their research skills. Gerry Allan and Alan Rimm-Kauffman allowed us to cite liberally
from their materials on customers and Internet marketing. We thank Valerie Redd and
Kelly Brandow for their help in designing, testing, and administering the survey of the
metrics that senior marketing managers use to monitor and manage their businesses.
Marc Goldstein combined business savvy with deft editing touches that improved the
readability of almost every chapter. Amy Neidlinger, Jeanne Levine, Andy Beaster,
Debbie Williams, Nonie Ratcliff, and their colleagues also made significant improvements in moving from a raw manuscript to the book in your hands.
Erv Shames, Erjen van Nierop, Peter Hedlund, Fred Telegdy, Judy Jordan, Lee Pielemier,
and Richard Johnson have collaborated on our “Allocator” management simulation
and “Management by the Numbers” online tutorials. That work helped us set the stage
for this volume. Finally, we thank Emily, Kate, Donna, and Sarah, who graciously tolerated the time sacrificed from home and social lives for the writing of this book.
For the third edition we would also like to thank Raymond Pirouz and Liz Gray for
sharing their opinions and expertise in respect to online marketing.
Thanks to Dr. Manuel Garcia-Garcia, Neuroscience Director, Nielsen and Adjunct
Associate Professor, NYU Stern School of Business, and Neuroscience UVA student
Pasha Davoudian for their invaluable guidance and assistance with the section on
neuro-marketing.
ix
This page intentionally left blank
ABOUT THE AUTHORS
Neil T. Bendle is an Assistant Professor of Marketing at the Ivey Business School,
Western University, Canada. He holds a PhD from the Carlson School of Management,
University of Minnesota, and an MBA from Darden. He has been published in journals such as Marketing Science and the Journal of Consumer Research. He has nearly a
decade’s experience in marketing management, consulting, business systems improvement, and financial management. He was responsible for measuring the success of
marketing campaigns for the British Labour Party.
Paul W. Farris is Landmark Communications Professor and Professor of Marketing
at The Darden Graduate Business School, University of Virginia, where he has taught
since 1980. Previously he was on the faculty of the Harvard Business School and worked
in marketing management for Unilever. Professor Farris’s research has produced
award-winning articles on retail power, the measurement of advertising effects, and
marketing budgeting. He has published many articles in journals such as the Harvard
Business Review, Journal of Marketing, Journal of Advertising Research, and Marketing
Science. He is currently developing improved techniques for integrating marketing
and financial metrics and is coauthor of several books, including The Profit Impact of
Marketing Strategy Project: Retrospect and Prospects. Farris’s consulting clients have
ranged from Apple and IBM to Procter & Gamble and Unilever. He has also served
on boards of manufacturers and retailers and as an academic trustee of the Marketing
Science Institute.
Phillip E. Pfeifer, Richard S. Reynolds Professor of Business Administration at The
Darden Graduate Business School, currently specializes in direct/interactive marketing. He has published a popular MBA textbook and more than 35 refereed articles in
journals such as the Journal of Interactive Marketing, Journal of Database Marketing,
Decision Sciences, and the Journal of Forecasting. In addition to academic articles and
a textbook, Mr. Pfeifer is a prolific case writer, having been recognized in 2004 as
the Darden School’s faculty leader in terms of external case sales, and in 2008 with a
Wachovia Award for Distinguished Case Writer. His teaching has won student awards
and has been recognized in Business Week’s Guide to the Best Business Schools. Recent
consulting clients include Circuit City, Procter & Gamble, and CarMax.
xi
Dr. David J. Reibstein is the William S. Woodside Professor and Professor of Marketing
at the Wharton School, University of Pennsylvania. Dave has been on the Wharton
Faculty for more than two decades. He was the Vice Dean of the Wharton School, and
Director of the Wharton Graduate Division. In 1999-2001, Dave took a leave of absence
from academia to serve as the Executive Director of the Marketing Science Institute.
He previously taught at Harvard, and was a Visiting Professor at Stanford, INSEAD,
and ISB (in India). Dave was the Chairman of the American Marketing Association.
He has a radio show, Measured Thoughts with Dave Reibstein, on SiriusXM Radio.
xii
MARKETING METRICS
FOREWORD
Despite its importance, marketing is one of the least understood, least measurable functions at many companies. With sales force costs, it accounts for 10 percent or more of
operating budgets at a wide range of public firms. Its effectiveness is fundamental to
stock market valuations, which often rest upon aggressive assumptions for customer
acquisition and organic growth. Nevertheless, many corporate boards lack the understanding to evaluate marketing strategies and expenditures. Most directors—and a rising percentage of Fortune 500 CEOs—lack deep experience in this field.
Marketing executives, for their part, often fail to develop the quantitative, analytical
skills needed to manage productivity. Right-brain thinkers may devise creative campaigns to drive sales but show little interest in the wider financial impact of their work.
Frequently, they resist being held accountable even for top-line performance, asserting
that factors beyond their control—including competition—make it difficult to monitor
the results of their programs.
In this context, marketing decisions are often made without the information, expertise,
and measurable feedback needed. As Procter & Gamble’s Chief Marketing Officer has
said, “Marketing is a $450 billion industry, and we are making decisions with less data
and discipline than we apply to $100,000 decisions in other aspects of our business.”
This is a troubling state of affairs. But it can change.
In a recent article in The Wall Street Journal, I called on marketing managers to take
concrete steps to correct it. I urged them to gather and analyze basic market data,
measure the core factors that drive their business models, analyze the profitability of
individual customer accounts, and optimize resource allocation among increasingly
fragmented media. These are analytical, data-intensive, left-brain practices. Going
forward, I believe they’ll be crucial to the success of marketing executives and their
employers. As I concluded in the Journal:
“Today’s boards want chief marketing officers who can speak the language of
productivity and return on investment and are willing to be held accountable. In
recent years, manufacturing, procurement and logistics have all tightened their
belts in the cause of improved productivity. As a result, marketing expenditures
account for a larger percentage of many corporate cost structures than ever
before. Today’s boards don’t need chief marketing officers who have creative flair
but no financial discipline. They need ambidextrous marketers who offer both.”
xiii
In Marketing Metrics, Farris, Bendle, Pfeifer, and Reibstein have given us a valuable
means toward this end. In a single volume, and with impressive clarity, they have
outlined the sources, strengths, and weaknesses of a broad array of marketing metrics.
They have explained how to harness those data for insight. Most importantly, they
have explained how to act on this insight—how to apply it not only in planning campaigns, but also in measuring their impact, correcting their courses, and optimizing
their results. In essence, Marketing Metrics is a key reference for managers who aim to
become skilled in both right- and left-brain marketing. I highly recommend it for all
ambidextrous marketers.
John A. Quelch, Lincoln Filene Professor of Business Administration and Senior
Associate Dean for International Development, Harvard Business School
xiv
MARKETING METRICS
FOREWORD TO THE
THIRD EDITION
At Google, we have a saying we use quite frequently: “Data beats opinion.” This mantra
inspires us to constantly think about how we can increase the ratio of fact to speculation. What do we actually know vs. what do we only think we know? The best approach
we’ve found is to determine our key performance indicators, and then measure how we
are doing against them on a regular basis. This allows us to optimize and expand those
programs that are working, while pulling back on those that are not.
In today’s hyper-competitive business landscape, most marketers are compelled to take
a similar approach. No longer can marketers rely on conventional wisdom, rules of
thumb, or intuition that may have been sufficient in the past.
The challenge, however, for all marketers is knowing what to measure and exactly how
to measure it.
That’s where Marketing Metrics comes in. In its first two editions, I’ve found it to be
the most comprehensive and authoritative guide to defining, constructing, and using
the metrics every marketer needs today. It’s a book I keep handy on my shelf and refer
to frequently.
As marketing continues to rapidly evolve, Marketing Metrics continues to stay at the
cutting edge. This third edition updates and adds more detail on a number of the key
metrics, including brand metrics and ROI. Given the increasing importance of online
and social metrics, this new edition now dedicates a chapter to them separate from traditional advertising metrics. Herein you will also find a section about the metrics for the
emerging area of neuro-marketing.
In our experience at Google, marketers who move with speed, make their messages
highly relevant, and use data (it beats opinion!) are best-positioned for success with
today’s buyers and modern media vehicles. I therefore heartily recommend Marketing
Metrics as the foundation of the data portion of this three-pronged recipe for marketing
success!
Jim Lecinski
Vice President, Americas Customer Solutions, Google
xv
This page intentionally left blank
1
INTRODUCTION
In recent years, data-based marketing has swept through the business world. In its
wake, measurable performance and accountability have become the keys to marketing success. However, few managers appreciate the range of metrics by which they can
evaluate marketing strategies and dynamics. Fewer still understand the pros, cons, and
nuances of each.
More than a decade ago, we recognized that marketers, general managers, and business
students needed a comprehensive, practical reference on the metrics used to judge marketing programs and quantify their results. This book was the result and seeks to provide that reference. This is now the third edition of the book and we continue to wish
our readers great success using this book to improve their understanding of marketing.
1.1 What Is a Metric?
A metric is a measuring system that quantifies a trend, dynamic, or characteristic.1
In virtually all disciplines, practitioners use metrics to explain phenomena, diagnose
causes, share findings, and project the results of future events. Throughout the worlds
of science, business, and government, metrics encourage rigor and objectivity. They
make it possible to compare observations across regions and time periods. They facilitate understanding and collaboration.
1.2 Why Do You Need Metrics?
“When you can measure what you are speaking about, and express it in numbers,
you know something about it; but when you cannot measure it, when you cannot
express it in numbers, your knowledge is of a meager and unsatisfactory kind: it may
be the beginning of knowledge, but you have scarcely, in your thoughts, advanced to
the stage of science.”—Lord Kelvin, Popular Lectures and Addresses (1891–94)2
1
Lord Kelvin, a British physicist and the manager in charge of laying the first successful
transatlantic cable, was one of history’s great advocates for quantitative investigation.
In his day, however, mathematical rigor had not yet spread widely beyond the worlds
of science, engineering, and finance. Much has changed since then.
Today, numerical fluency is a crucial skill for every business leader. Managers must
quantify market opportunities and competitive threats. They must justify the financial
risks and benefits of their decisions. They must evaluate plans, explain variances, judge
performance, and identify leverage points for improvement—all in numeric terms.
These responsibilities require a strong command of measurements and of the systems
and formulas that generate them. In short, they require metrics.
Managers must select, calculate, and explain key business metrics. They must understand how each is constructed and how to use it in decision-making. Witness the following, more recent quotes from management experts:
“. . . every metric, whether it is used explicitly to influence behavior, to evaluate
future strategies, or simply to take stock, will affect actions and decisions.” 3
“If you can’t measure it, you can’t manage it.”4
1.3 Marketing Metrics: Opportunities, Performance,
and Accountability
Marketers are by no means immune to the drive toward quantitative planning and
evaluation. Marketing may once have been regarded as more an art than a science.
Executives may once have cheerfully admitted that they knew they wasted half the
money they spent on advertising, but they didn’t know which half. Those days, however, are gone.
Today, marketers must understand their addressable markets quantitatively. They
must measure new opportunities and the investment needed to realize them. Marketers
must quantify the value of products, customers, and distribution channels—all under
various pricing and promotional scenarios. Increasingly, marketers are held accountable for the financial ramifications of their decisions. Observers have noted this trend in
graphic terms:
“For years, corporate marketers have walked into budget meetings like neighborhood
junkies. They couldn’t always justify how well they spent past handouts or what difference it all made. They just wanted more money—for flashy TV ads, for big-ticket
events, for, you know, getting out the message and building up the brand. But those
heady days of blind budget increases are fast being replaced with a new mantra:
measurement and accountability.”5
2
MARKETING METRICS
1.4 Choosing the Right Numbers
The numeric imperative represents a challenge, however. In business and economics,
many metrics are complex and difficult to master. Some are highly specialized and
best suited to specific analyses. Many require data that may be approximate, incomplete, or unavailable.
Under these circumstances, no single metric is likely to be perfect. For this reason, we
recommend that marketers use a portfolio or “dashboard” of metrics. By doing so,
they can view market dynamics from various perspectives and arrive at “triangulated”
strategies and solutions. Additionally, with multiple metrics, marketers can use each as
a check on the others. In this way, they can maximize the accuracy of their knowledge.6
They can also estimate or project one data point on the basis of others. Of course, to use
multiple metrics effectively, marketers must appreciate the relations between them and
the limitations inherent in each.
When this understanding is achieved, however, metrics can help a firm maintain a
productive focus on customers and markets. They can help managers identify the
strengths and weaknesses in both strategies and execution. Mathematically defined
and widely disseminated, metrics can become part of a precise, operational language
within a firm.
Data Availability and Globalization of Metrics
A further challenge in metrics stems from wide variations in the availability of
data between industries and geographies. Recognizing these variations, we have tried
to suggest alternative sources and procedures for estimating some of the metrics in
this book.
Fortunately, although both the range and type of marketing metrics may vary between
countries,7 these differences are shrinking rapidly. Ambler,8 for example, reports that
performance metrics have become a common language among marketers, and that they
are now used to rally teams and benchmark efforts internationally.
1.5 What Are We Measuring?
Measuring marketing is highly challenging. For example, marketers generally agree that
a firm’s brand is a key marketing asset but different marketers all have subtly different
views of what is meant by a brand. It is hard to measure something when you don’t
know what exactly you are trying to measure. We, therefore, suggest that the first thing
a marketer needs to establish is a clear definition of what they are trying to measure.
Chapter 1 Introduction
3
Watt and van den Berg distinguish theoretical and operations definitions in a way that
we find useful.
“Concepts represent the “real world” phenomena being explained by the theory. The
scientific method requires that the nature of these concepts be unambiguously communicated to others. This requirement mandates the creation of theoretical definitions….. Concepts must also be objectively observed. This requires that we create
operational definitions, which translate the verbal concepts into corresponding variables which can be measured.” Watt and van den Berg (p. 11)
The same authors differentiate constructs from concepts, arguing that the former are
even more abstract than concepts and cannot be directly observed. They use “source
credibility” as an example of a construct that comprises concepts such as expertise,
status, and objectivity. Of course constructs can also be operationalized in a number of
ways.
To see what this means, note that marketing has a number of basic ideas that capture
real world phenomena; let us call these concepts. These basic ideas are very important
to marketers and can be explained, and even formally defined, verbally. These concepts
are, however, not the same as metrics. For example, loyalty is a critical concept for
many marketers but my idea of loyalty may differ from yours. Is loyalty demonstrated
when I visit a grocery store every week? What if that grocery store is the only one I can
easily get to? I might not feel loyal to the store but I still visit it every week. Someone else
might feel highly loyal to the same store but live much further away and only be able to
visit irregularly. Which, if any, of these consumers are loyal?
We must make concrete our abstract concept of loyalty by providing an operational
definition, a precise specification in numerical terms of what exactly we mean. This
allows us to create metrics to keep track of how a firm is performing against the operational definitions specified. This book aims to improve measurement validity, how well
you translate your ideas into numbers; we do not seek to provide new ways of looking
at marketing, or argue which concepts are more important than any others.
Some common ways of translating concepts into metrics are shown below in Table 1.1.
Table 1.1 Common Metrics Used to Track Important Concepts
Concept
Metric(s)
Loyalty
Share of Requirements (SOR), Willingness to Pay (WTP)
Distribution
All Commodity Volume (ACV), Total ACV
Market Concentration
Three-firm concentration ratio, Herfindahl Index
4
MARKETING METRICS
Note in Table 1.1 that willingness to pay (WTP) measures both a consumer’s loyalty in
the sense of attachment and differentiation, the fact that the product is very different
and so there isn’t a close competitor that a consumer can purchase from instead. It is
important to consider what exactly you are measuring and what metric fits your precise
definition best.
Keeping a clear distinction between concepts, operational definitions, and metrics is
surprisingly hard. In any given marketing team or organization, one can expect to see a
certain level of confusion. We hope our book helps reduce this confusion and promote
a common language, but we are realists. Indeed we are happy to acknowledge that we
also make mistakes and inadvertently refer to metrics by the name of the concept. We
are trying to be clear but please let us know if you see areas where we can improve (just
in case there is a fourth edition).
There will continue to be healthy (or at least vigorous) debates in marketing on what
should be meant by various theoretical concepts and constructs. However, at the level
of measurement and reporting we believe that the field should be striving for consistency, accuracy, and reliability that allows us to at least understand what other people
mean, even if we disagree with what they are suggesting. No shared understanding can
happen without clear operational definitions. Providing these definitions is the primary
focus of the Common Language Project. This project to improve the measurement of
marketing, specifically making measurement in the discipline more consistent, is being
undertaken by MASB (the Marketing Accountability Standards Board, http://www.
themasb.org/), along with MSI (the Marketing Science Institute, http://www.msi.org/),
ANA (the Association of National Advertisers, https://www.ana.net/), and AMA (the
American Marketing Association, https://www.ama.org). We encourage readers to
learn more about, and support, the initiative.
1.6 Value of Information
There exists almost an infinite number of metrics that could be calculated. Even the
most quantitative marketer will recognize that more calculations don’t always help
make better decisions. Thus one question a marketer may want to start with is: “when
is a metric useful?”
A classic distinction is between data versus information versus knowledge. Data is what
we have a profusion of in the world of big data. Data is in raw form and doesn’t tell us
anything without being manipulated in some way. Information is data that has been
converted into something that can be used by a human reader. Ideally, information
gets converted into knowledge when a user understands and internalizes the information. Thus one way of thinking about the value of information is whether it creates
knowledge or not. Data that is simply being stored is not currently valuable, but often
has the potential to be valuable if approached in the right way. How can we extract the
Chapter 1 Introduction
5
information from the data we have? (Clearly the marketer should ensure they have
the legal and ethical right to use the data in this manner. Consent is usually a key consideration, but discussing law and ethics is beyond the scope of our book. The Direct
Marketing Association has resources to aid such thinking, http://thedma.org/.)
One way to increase the value of information is thus to make it easier for users to
convert it to knowledge. To do this we recommend considering how the information
you have extracted, such as the metrics you have calculated, can be presented in a userfriendly way. There are now many companies, e.g., Tableau software, www.tableau.
com, that specialize in translating information into visual representations. Such visual
depictions are an excellent aid to allowing the user to more easily extract the message
from the information you provide them.
An alternative way of thinking about the value of information is whether the information helps us take an action. Information is valuable only if it allows us to make a better decision. To cast this in terms of metrics, a metric’s value arises from its ability to
improve our decisions in some way. Note this is a very pragmatic approach as the value
of the metric depends upon what the user can do with the result. A CMO might find
estimates of the value of the brand she controls invaluable when arguing for increasing
the marketing budget with her C-suite colleagues. A more junior marketer, however,
may feel that he can’t impact brand value in any significant way so knowing this number is of no value to him. The more junior marketer can, however, impact whether the
product is on the retailers’ shelves and so may find distribution measures invaluable.
A related point is that people sometimes equate the value of information with the range
of possible alternatives that the metric can take. Knowing the precise number for a
metric that swings wildly can be very informative and thus valuable. If the metric never
changes significantly, knowing its precise reading at any given point is unlikely to be
very valuable. For example, information on the sales of a fashion item where consumer
reaction is unpredictable can be exceptionally valuable for stock planning. Estimates for
items with more predictable sales, e.g., safety matches, are less valuable because knowing the precise sales number is less likely to change the inventory order you would make
without the more refined sales estimate. For items with very stable sales, your estimate
based upon last year is likely to be good regardless of whether you calculate the precise
metric for this year.
Testing is a critical component of marketing plans, but where should you spend
your testing budget? What gives you the most information for your money? Scott
Armstrong notes that this depends upon what you are trying to achieve. Sometimes you
will want to emulate much academic research and drill down into a very specific topic.
This can lead to very consistent estimates, also known as being “reliable.” This means
every time you measure you get a similar result because you measure exactly the same
thing each time you measure. In everyday life the electronic scale that weighs you every
morning is reliable, you generally get the same result if nothing changes. This approach
6
MARKETING METRICS
makes sense if it is critical for you to be very precise, if small changes in a metric would
radically alter your plans.
More often, however, you aren’t sure you are measuring the right thing. You want to
know how the firm is performing generally but you have a less than perfect understanding of what performance means exactly. You might be interested in your general health
rather than your precise weight. Your weight is likely to be connected to your general
health but is far from the complete picture. In such situations you are interested in
whether the measures you are using are valid, whether the measures accurately capture
what you want them to capture. To assess validity you are likely to want multiple measures, in which case you’ll spread your testing budget across a wider range of tests and
are more tolerant of conflicting results. To assess your health you might look at your
weight, your blood pressure, your blood sugar, the ease of your breathing, etc. These
will sometimes point in different directions but put together they give a more comprehensive picture than fixating upon a single metric, however reliably the single metric
can be measured.
To have valid estimates of hard-to-define concepts, such as performance, we often
recommend a variety of tests and the use of multiple metrics. As we will discuss in
Chapters 12 and 13 it is often possible to have one metric look very good while the true
value of the company is destroyed. Testing multiple different areas and assessing different metrics may give you a less precise picture (it is less reliable) but is much less likely
to miss a major problem (it is more valid).
1.7 Mastering Metrics
Being able to “crunch the numbers” is vital to success in marketing. Knowing which
numbers to crunch, however, is a skill that develops over time. Toward that end,
managers must practice the use of metrics and learn from their mistakes. By working
through the examples in this book, we hope our readers will gain both confidence and
a firm understanding of the fundamentals of data-based marketing. With time and
experience, we trust that you will also develop an intuition about metrics, and learn to
dig deeper when calculations appear suspect or puzzling.
Ultimately, with regard to metrics, we believe many of our readers will require not
only familiarity but also fluency. That is, managers should be able to perform relevant calculations on the fly—under pressure, in board meetings, and during strategic deliberations and negotiations. Although not all readers will require that level of
fluency, we believe it will be increasingly expected of candidates for senior management positions, especially those with significant financial responsibility. We anticipate that a mastery of data-based marketing will become a means for many of our
Chapter 1 Introduction
7
readers to differentiate and position themselves for career advancement in an ever more
challenging environment.
1.8
Where are the “Top Ten” Metrics?
Working on this book we received many requests to provide a short list of the “key” or
“top ten” marketing metrics. The intuition behind this request is that readers (managers
and students) want to be able to focus their attention on the “most important” metrics.
Although some readers have read the earlier editions from cover to cover, it is safe to
say none of the authors have had that pleasure. We view the book as a reference book—
something to keep on the shelf and use when confronted with a new or less familiar
metric. The list of metrics covered is therefore long so as to be useful for those occasions. It is not aimed to be a guide to the X number of metrics you must apply to monitor marketing. It is this view of the book as a reference guide that helps explain why we
do not rate or rank the long list of metrics. We see you pulling the book from the shelf
as needed, rather than us pushing our preferred metrics upon you.
Specifically, the reasons for us not providing the short list of “really important” metrics
are as follows.
First, we believe that any ranking of marketing metrics from most to least useful should
depend on the type of business under consideration. Thus what metrics you prefer
depend upon what you need them for. For example, marketers of business-to-business
products and services that go to market through a direct sales force don’t need metrics
that measure retail availability or dealer productivity.
Also, even what might begin as a short list tends to expand rapidly as metrics come
in matched sets. For example, if customer lifetime value is important to your business
(let’s say, financial services), then you are also likely to use measures of retention and
acquisition costs. The same notion applies to retail, media, sales force, and Web traffic
metrics. If some of these are important to you, others in the same general categories are
likely to be rated as useful, too.
Third, businesses don’t always have access (at a reasonable cost) to the metrics they
would like to have. Inevitably, some of the rankings presented will reflect the cost of
obtaining the data that underlie the particular metrics. Some metrics may be interesting
to know but cost more to obtain than the value of the information they provide. The
size of the organization will thus matter. Small organizations will use metrics that are
cheaper to obtain; whereas larger organizations are more likely to be able to realize the
full value from expensive, proprietary, or custom-created metrics. The same goes for
stages in the product life cycle. Managers of newly launched products often have different concerns and metrics to monitor them than mature products.
8
MARKETING METRICS
Fourth, we believe that some metrics currently ranked lower by managers will ultimately prove to be very useful, after managers fully understand the pros and cons of a
particular metric. For example, advocates believe that Economic Value Added (EVA) is
the “gold standard” of profitability metrics, but when we discuss it with many managers, it ranks far below other financial performance measures such as ROI. We believe
one reason for the low ranking of EVA is that this metric is less applicable at the “operating level” than for overall corporate performance. So even within the same business,
depending on where a manager sits in the organization, some metrics are more relevant
than others. Also, like EVA, many metrics that we have included are relatively new to
marketing, and many managers don’t understand them well or know how they might
be relevant to their particular business. Customer Lifetime Value is another metric that
is gaining acceptance, but is still unfamiliar to many managers. If all these metrics were
perfectly understood, there would be no need for a book of this type.
We included the results of our survey of marketing managers in the second edition so
that readers could learn what metrics other managers thought were potentially useful.
However, we are now less convinced that the survey results are useful, because metric
use and understanding remains an awfully long way from where we want to be. In this
third edition we have therefore concentrated on adding more metrics and explaining
some metrics in greater detail. Deleting the managerial survey results from the second
edition gave us the space to do this.
Here we simply note the key points from the survey. These were that managers value
the profit related metrics, net profit, ROI, and margin most highly, even though these
metrics had less to do with day-to-day marketing decisions. We presume this is because
those are the metrics they are asked about by those who control budgets. Customer
satisfaction was the most popular “non-financial” metric. Sales related metrics, such as
Sales Total also proved popular.
1.9 What’s New in the Third Edition?
We have used the space gained in removing the survey to go deeper into areas that have
risen in importance in the years since the second edition. The biggest change is our new
chapter on online metrics, which seeks to clarify an area where there is no shortage of
data and a profusion of measures. Unfortunately it remains an area in which it can be
extremely difficult for a manager to understand what is happening often due to lack of
standardization of definitions. We have also added a new section on neuro-marketing
to allow managers to gain some familiarity with exciting developments in this field.
Other significant changes include adding more detail on brand valuation, advertising
elasticity, and clarifying the concept of double jeopardy.
Chapter 1 Introduction
9
New metrics are particularly in need of a careful appraisal with respect to their reliability and validity as metrics to inform management decisions. This edition expands our
discussion of these concepts as well.
A closely related concept is “the value of information”—e.g., is it worth doing more
market research or testing?—which is a critical input to many managerial decisions.
As such we have also added more detail on the value of information and how managers can use estimates of its value to make better informed decisions. We have added a
discussion of the Gross Model for budget allocation and outlined our own variant of
the budgeting model which describes how a manager can decide when to stop creating and testing further advertising copy and roll out the current best candidate. The
space devoted to this model for advertising copy and media funds is not because this
particular decision is more important than other decisions, but because it illustrates a
general approach to thinking about how using marketing metrics can improve marketing decisions.
We hope that you enjoy the new edition of Marketing Metrics.
Organization of the Text
This book is organized into chapters that correspond to the various roles played by
marketing metrics in enterprise management. Individual chapters are dedicated to
metrics used in promotional strategy, advertising, and distribution, for example. Each
chapter is composed of sections devoted to specific concepts and calculations.
We must present these metrics in a sequence that will appear somewhat arbitrary. In
organizing this text, we have sought to strike a balance between two goals: (1) to establish core concepts first and build gradually toward increasing sophistication, and (2)
to group related metrics in clusters, helping our readers recognize patterns of mutual
reinforcement and interdependence. In Figure 1.1, we offer a graphical presentation of
this structure, demonstrating the interlocking nature of all marketing metrics—indeed
of all marketing programs—as well as the central role of the customer.
The central issues addressed by the metrics in this book are as follows:
■
■
■
■
10
Chapter 2—Share of Hearts, Minds, and Markets: Customer perceptions, market share, and competitive analysis.
Chapter 3—Margins and Profits: Revenues, cost structures, and profitability.
Chapter 4—Product and Portfolio Management: The metrics behind product
strategy, including measures of trial, growth, cannibalization, and brand equity.
Chapter 5—Customer Profitability: The value of individual customers and
relationships.
MARKETING METRICS
Customers and Market Research
Logistics
Customer
Profitability
Operations
Product and
Portfolio
Management
Margins and
Profits
Finance
Sales Force
Sales Force
and
Channel
Management
Share of Hearts, Minds,
and Markets
Marketing
and
Finance
Pricing
Strategy
Promotions
The Trade
Online, Email,
and Mobile
Metrics
Advertising
Metrics
Advertising
Agency
Figure 1.1
■
■
■
■
■
Marketing Metrics: Marketing at the Core of the Organization
Chapter 6—Sales Force and Channel Management: Sales force organization,
performance, and compensation. Distribution coverage and logistics.
Chapter 7—Pricing Strategy: Price sensitivity and optimization, with an eye
toward setting prices to maximize profits.
Chapter 8—Promotion: Temporary price promotions, coupons, rebates, and
trade allowances.
Chapter 9—Advertising Metrics: The central measures of advertising coverage
and effectiveness, including reach, frequency, rating points, and impressions.
Models for consumer response to advertising.
Chapter 10—Online, Email, and Mobile Metrics: Specialized metrics for Webbased, Mobile, and Email campaigns.
Chapter 1 Introduction
11
■
■
■
Chapter 11—Marketing and Finance: Financial evaluation of marketing
programs.
Chapter 12—The Marketing Metrics X-Ray: The use of metrics as leading indicators of opportunities, challenges, and financial performance.
Chapter 13—System of Metrics: Decomposing marketing metrics into component parts can improve measurement accuracy, add managerial insight into
problems, and assist marketing model building.
Components of Each Chapter
As shown in Table 1.2, the chapters are composed of multiple sections, each dedicated
to specific marketing concepts or metrics. Within each section, we open with definitions, formulas, and a brief description of the metrics covered. Next, in a passage titled
Construction, we explore the issues surrounding these metrics, including their formulation, application, interpretation, and strategic ramifications. We provide examples
to illustrate calculations, reinforce concepts, and help readers verify their understanding of key formulas. That done, in a passage titled Data Sources, Complications, and
Cautions, we probe the limitations of the metrics under consideration and potential
pitfalls in their use. Toward that end, we also examine the assumptions underlying
these metrics. Finally, we close each section with a brief survey of Related Metrics and
Concepts.
In organizing the text in this way, our goal is straightforward: Most of the metrics in
this book have broad implications and multiple layers of interpretation. Doctoral theses
could be devoted to many of them, and have been written about some. In this book,
however, we want to offer an accessible, practical reference. If the devil is in the details,
we want to identify, locate, and warn readers against him, but not to elaborate his
entire demonology. Consequently, we discuss each metric in stages, working progressively toward increasing levels of sophistication. We invite our readers to sample this
information as they see fit, exploring each metric to the depth that they find most useful
and rewarding.
With an eye toward accessibility, we have also avoided advanced mathematical notation. Most of the calculations in this book can be performed by hand, on the back of the
proverbial envelope. More complex or intensive computations may require a spreadsheet. Nothing further should be needed.
12
MARKETING METRICS
Table 1.2 Major Metrics List
Section
Metric
Share of Hearts, Minds, and Markets
Section
Metric
3.2
3.3
3.3
3.4
3.5
3.6
3.6
3.6
3.7
3.7
Channel Margins
Average Price per Unit
Price Per Statistical Unit
Variable and Fixed Costs
Marketing Spending
Contribution per Unit
Contribution Margin (%)
Break-Even Sales
Target Volume
Target Revenues
2.1
Revenue Market Share
2.1
Unit Market Share
2.2
Relative Market Share
2.3
Brand Development Index
2.3
Category Development
Index
2.4–2.6
Decomposition of Market
Share
2.4
Market Penetration
2.4
Brand Penetration
2.4
Penetration Share
2.5
Share of Requirements
2.6
Usage Index
2.7
Hierarchy of Effects
2.7
Awareness
2.7
Top of Mind
2.7
Ad Awareness
2.7
Knowledge
2.7
Consumer Beliefs
2.7
Purchase Intentions
2.7
Purchase Habits
2.7
Loyalty
2.7
Likeability
2.8
Willingness to Recommend
2.8
Customer Satisfaction
Customer Profitability
2.9
Net Promoter
2.10
Willingness to Search
2.11
Neuro-Marketing
5.1
5.1
5.1
5.2
5.3
5.4
Margins and Profits
3.1
Unit Margin
3.1
Margin (%)
Product and Portfolio Management
4.1
4.1
4.1
4.1
4.2
4.2
4.3
4.3
4.4
4.5
4.6
4.7
Trial
Repeat Volume
Penetration
Volume Projections
Year-on-Year Growth
Compound Annual Growth
Rate (CAGR)
Cannibalization Rate
Fair Share Draw Rate
Brand Equity Metrics
Conjoint Utilities
Segment Utilities
Conjoint Utilities and
Volume Projections
Customers
Recency
Retention Rate
Customer Profit
Customer Lifetime Value
Prospect Lifetime Value
Continues
Chapter 1 Introduction
13
Table 1.2
Continued
Section
Metric
Section
Metric
5.5
Average Acquisition Cost
7.4
Optimal Price
5.5
Average Retention Cost
7.5
Residual Elasticity
Sales Force and Channel Management
Promotion
6.1
Workload
8.1
Baseline Sales
6.1
Sales Potential Forecast
8.1
6.2
Sales Goal
Incremental Sales/
Promotion Lift
6.3
Sales Force Effectiveness
8.2
Redemption Rates
6.4
Compensation
8.2
6.4
Break-Even Number of
Employees
Costs for Coupons and
Rebates
8.2
6.5
Sales Funnel, Sales Pipeline
Percentage Sales with
Coupon
6.6
Numeric Distribution
8.3
Percent Sales on Deal
6.6
All Commodity Volume
(ACV)
8.3
Pass-Through
8.4
Price Waterfall
6.6
Product Category Volume
(PCV)
6.6
Total Distribution
6.6
Category Performance Ratio
6.7
Out of Stock
6.7
Advertising Metrics
9.1
Impressions
9.1
Gross Rating Points (GRPs)
9.2
Cost per Thousand
Impressions (CPM)
Inventories
9.3
Net Reach
6.8
Markdowns
9.3
Average Frequency
6.8
Direct Product Profitability
(DPP)
9.4
Frequency Response
Functions
6.8
Gross Margin Return on
Inventory Investment
(GMROII)
9.5
Effective Reach
9.5
Effective Frequency
9.6
Share of Voice
9.7
Advertising Elasticity of
Demand
Pricing Strategy
7.1
Price Premium
7.2
Reservation Price
7.2
Percent Good Value
7.3
Price Elasticity of Demand
14
MARKETING METRICS
Table 1.2 Continued
Section
Metric
Section Metric
Online, Email, and Mobile Metrics
Marketing and Finance
10.1
Pageviews
11.1
Net Profit
10.2
Rich Media Display Time
11.1
Return on Sales (ROS)
10.2
Rich Media Interaction Rate
11.1
10.3
Clickthrough Rate
10.4
Cost per Click
Earnings Before Interest, Taxes,
Depreciation, and Amortization
(EBITDA)
10.4
Cost per Order
11.2
Return on Investment (ROI)
10.4
Cost per Customer Acquired
11.3
Economic Profit (aka EVA®)
10.5
Visits
11.4
Payback
10.5
Visitors
11.4
Net Present Value (NPV)
10.5
Abandonment Rate
10.6
Bounce Rate (Web site)
11.4
11.5
Internal Rate of Return (IRR)
Marketing Return on Investment
(MROI)
10.7
Friends/Followers/ Supporters
10.7
Likes
10.7
Value of a Like
10.8
Downloads
10.9
Average Revenue per User
10.10
Email Metrics
Reference Materials
Throughout this text, we have highlighted formulas and definitions for easy reference. We have also included outlines of key terms at the beginning of each chapter
and section. Within each formula, we have followed this notation to define all inputs
and outputs.
$—(Dollar Terms): A monetary value. We have used the dollar sign and “dollar terms” for brevity, but any other currency, including the euro, yen, dinar, or
yuan, would be equally appropriate.
%—(Percentage): Used as the equivalent of fractions or decimals. For readability, we have intentionally omitted the step of multiplying decimals by 100 to
obtain percentages.
#—(Count): Used for such measures as unit sales or number of competitors.
Chapter 1 Introduction
15
R—(Rating): Expressed on a scale that translates qualitative judgments or preferences into numeric ratings. Example: A survey in which customers are asked
to assign a rating of “1” to items that they find least satisfactory and “5” to those
that are most satisfactory. Ratings have no intrinsic meaning without reference to
their scale and context.
I—(Index): A comparative figure, often linked to or expressive of a market
average. Example: the consumer price index. Indexes are often interpreted as
a percentage.
$—Dollar. %—Percentage. #—Count. R—Rating. I—Index.
References and Suggested Further Reading
Abela, Andrew, Bruce H. Clark, and Tim Ambler. “Marketing Performance Measurement,
Performance, and Learning,” working paper, September 1, 2004.
Ambler, Tim, and Chris Styles. (1995). “Brand Equity: Toward Measures That Matter,” working
paper No. 95-902, London Business School, Centre for Marketing.
Armstrong, J. Scott (1974) “Eclectic Research and Construct Validation,” Published in Models of
Buyer Behavior: Conceptual, Quantative, and Empirical, Jagdish N. Sheth (Editor), (pages 3-14),
Harper and Row, New York, NY.
Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used and Where?”
Marketing Science Institute, (03-111), working paper, Series issues two 03-002.
Clark, Bruce H., Andrew V. Abela, and Tim Ambler. “Return on Measurement: Relating
Marketing Metrics Practices to Strategic Performance,” working paper, January 12, 2004.
Hauser, John, and Gerald Katz. (1998). “Metrics: You Are What You Measure,” European
Management Journal, Vo. 16, No. 5, pp. 517–528.
Kaplan, R. S., and D. P. Norton. (1996). The Balanced Scorecard: Translating Strategy into
Action, Boston, MA: Harvard Business School Press.
Watt, James H., and Sjef van den Berg, (1995), Research Methods for Communication Science,
Allyn & Bacon.
16
MARKETING METRICS
2
SHARE OF HEARTS, MINDS,
AND MARKETS
Introduction
Key concepts covered in this chapter:
Market Share
Usage Index
Relative Market Share
Awareness, Attitudes, and Usage (AAU)
Market Concentration
Customer Satisfaction
Brand Development Index (BDI)
Willingness to Recommend
Category Development Index (CDI)
Net Promoter
Penetration
Willingness to Search
Share of Requirements
Neuro-marketing
“As Wal-Mart aggressively rolls out more stores, it continues to capture an increasing share of wallet. Three out of five consumers shopped for gifts at Wal-Mart this
past holiday season. U.S. households now buy, on average, 22% of their groceries at
Wal-Mart. A quarter of all shoppers indicate that they are spending more of their
clothing budget at Wal-Mart now compared with a year ago. These ShopperScape
findings lend credence to Retail Forward’s premise that Wal-Mart will continue to
push the boundaries of what consumers will allow it to be.”1
17
At first glance, market share appears to involve a relatively simple calculation: “us/
(us + them).” But this raises a host of questions. Who, for example, are “they?” That is,
how broadly do we define our competitive universe? Which units are used? Where in
the value chain do we capture our information? What time frame will maximize our
signal-to-noise ratio? In a metric as important as market share, and in one as closely
monitored for changes and trends, the answers to such questions are crucial. In this
chapter, we will address them and also introduce key components of market share,
including penetration share, usage index, and share of requirements.
Probing the dynamics behind market share, we’ll explore measures of awareness, attitude, and usage—major factors in the decision-making process by which customers
select one brand over another. We’ll discuss customer satisfaction with products and
dealers, the quantification of which is growing in importance among marketing professionals. Finally, we’ll consider metrics measuring the depth of consumer preference
and satisfaction, including customers’ willingness to search if a brand is unavailable and
their disposition to recommend that brand to others. Increasingly, marketers rely on
these as leading indicators of future changes in share.
Metric
Construction
Considerations
Revenue Market
Share
Sales revenue as
a percentage of
market sales
revenue.
Scope of market
Measure of
definition. Channel competitiveness.
level analyzed.
Before/ after
discounts. Time
period covered.
2.1
Unit Market Share
Unit sales as a
percentage of
market unit sales.
Scope of market
Measure of
definition. Channel competitiveness.
level analyzed.
Time period
covered.
2.2
Relative Market
Share
Brand market
Can use either unit Assesses
share divided by
or revenue shares. comparative market
largest competitor’s
strength.
market share.
2.1
18
MARKETING METRICS
Purpose
Metric
Construction
Considerations
2.3
Brand
Development
Index
Brand sales in a
specified segment,
compared with
sales of that brand
in the market as a
whole.
Can use either unit Regional or
or revenue sales.
segment differences
in brand purchases
and consumption.
2.3
Category
Development
Index
Category sales
Can use either unit
in a specified
or revenue sales.
segment, compared
with sales of that
category in the
market as a whole.
Regional or
segment differences
in category
purchases and
consumption.
2.4
Decomposition
of Market Share
Penetration
Share * Share of
Requirements
* Usage Index.
Time period
covered.
Calculation of
market share.
Competitive
analysis. Historical
trends analysis.
Formulation
of marketing
objectives.
2.4
Category
Penetration
Purchasers of a
product category
as a percentage of
total population.
Based on
population.
Therefore, unit/
revenue
consideration not
relevant.
Measures category
acceptance by a
defined population.
Useful in tracking
acceptance of new
product categories.
2.4
Brand Penetration
Purchasers of
a brand as a
percentage of total
population.
Based on
population.
Therefore,
unit/revenue
consideration not
relevant.
Measures brand
acceptance by
a defined
population.
2.5
2.6
Purpose
Continues
Chapter 2 Share of Hearts, Minds, and Markets
19
Metric
Construction
Considerations
Purpose
2.4
Penetration Share
The ratio of
brand penetration
to category
penetration.
Also the percentage
of category buyers
who bought
the brand. A
component of
the market share
formula.
Comparative
acceptance of the
brand within its
category.
2.5
Share of
Requirements
Brand purchases
as a percentage
of total category
purchases by
buyers of that
brand.
Purchases can be
either units or
revenues.
Level of
commitment to
a brand by its
existing customers.
2.6
Usage Index
The ratio of
Can be used
average category
with both unit or
purchases of
revenue sales.
customers of the
brand to the overall
average category
purchases per
category customer.
Measures relative
usage of a
category by
customers of a
specific brand.
2.7
Hierarchy of
Effects
Awareness;
attitudes, beliefs;
importance;
intentions to try;
buy; trial, repeat.
Strict sequence is
often violated and
can be reversed.
Set marketing
and advertising
objectives.
Understand
progress in stages
of customer
decision process.
2.7
Awareness
Percentage of total
population that is
aware of a brand.
Is this prompted
or unprompted
awareness?
Consideration of
who has heard of
the brand.
2.7
Top of Mind
First brand to
consider.
May be subject
to most recent
advertising or
experience.
Saliency of brand.
20
MARKETING METRICS
Metric
Construction
Considerations
Purpose
2.7
Ad Awareness
Percentage of total
population that is
aware
of a brand’s
advertising.
May vary by
schedule, reach,
and frequency of
advertising.
One measure
of advertising
effects. May
indicate “stopping
power” of ads.
2.7
Knowledge
Percentage of
population
with knowledge
of product,
recollection of its
advertising.
Not a formal
metric. Is this
prompted or
unprompted
knowledge?
Extent of
familiarity with
product beyond
name recognition.
2.7
Consumer
Beliefs
Customers/
consumers view of
product, generally
captured via survey
responses, often
through
ratings on a
scale.
Customers/
consumers may
hold beliefs with
varying degrees of
conviction.
Perception of
brand by attribute.
2.7
Purchase
Intentions
Probability of
intention to
purchase.
To estimate
probability of
purchase, aggregate
and analyze ratings
of stated intentions
(for example, top
two boxes).
Measures
pre-shopping
disposition to
purchase.
2.7
Purchase Habits
Frequency of
May vary widely
purchase. Quantity among shopping
typically purchased. trips.
2.7
Loyalty
Measures
include share of
requirements,
willingness to
pay premium,
willingness to
search.
Helps identify
heavy users.
“Loyalty” itself
Indication of base
is not a formal
future revenue
metric, but specific stream.
metrics measure
aspects of this
dynamic. New
product entries
may alter loyalty
levels.
Continues
Chapter 2 Share of Hearts, Minds, and Markets
21
Metric
Construction
2.7
Likeability
Generally measured Often believed
via ratings across a to correlate with
number of scales.
persuasion.
Shows overall
preference prior to
shopping.
2.8
Willingness to
Recommend
Generally measured Nonlinear in
via ratings across a impact.
1–5 scale.
Shows strength of
loyalty, potential
impact on others.
2.8
Customer
Satisfaction
Generally measured
on a 1–5 scale, in
which customers
declare their
satisfaction with
brand in general or
specific attributes.
Subject to response
bias. Captures
views
of current
customers, not
lost customers.
Satisfaction is a
function of
expectations.
Indicates
likelihood of
repurchase.
Reports of
dissatisfaction show
aspects that require
improvement to
enhance loyalty.
2.9
Net Promoter
Percentage of
customers willing
to recommend
to others less
the percentage
unwilling to
recommend the
product or service.
Requires a survey
of intentions.
Some claim it
to be the single
best metric for
marketers.
2.10
Willingness to
Search
Percentage of
customers willing
to delay purchases,
change stores, or
reduce quantities
to avoid switching
brands.
Hard to capture.
Indicates
importance of
distribution
coverage.
2.11
Neuro-Marketing
Measures
Technological
advances have
allowed marketers
to gain greater
insight into how
consumers think.
Requires expertise
and sometimes
expensive
equipment.
Can help illuminate
reactions that
consumers find
hard to verbalize.
22
MARKETING METRICS
Considerations
Purpose
2.1 Market Share
Market share is the percentage of a market (defined in terms of either units or revenue) accounted for by a specific entity.
Unit Sales (#)
Unit Market Share (%) =
Total Market Unit Sales (#)
Revenue Market Share (%) =
Sales Revenue ($)
Total Market Revenue ($)
Unit Market Share is also known as Volume Share. Revenue Market Share is also
known as Value Share. Marketers need to be able to translate sales targets into market
share because this will demonstrate whether forecasts are to be attained by growing
with the market or by capturing share from competitors. The latter will almost always
be more difficult to achieve. Market share is closely monitored for signs of change in
the competitive landscape, and it frequently drives strategic or tactical action.
Purpose: Key indicator of market competitiveness.
Market share is an indicator of how well a firm is doing against its competitors. This
metric, supplemented by changes in sales revenue, helps managers evaluate both primary and selective demand in their market. That is, it enables them to judge not only
total market growth or decline but also trends in customers’ selections among competitors. Generally, sales growth resulting from primary demand (total market growth) is
less costly and more profitable than that achieved by capturing share from competitors.
Conversely, losses in market share can signal serious long-term problems that require
strategic adjustments. Firms with market shares below a certain level may not be viable.
Similarly, within a firm’s product line, market share trends for individual products are
considered early indicators of future opportunities or problems.
Construction
Market Share: The percentage of a market accounted for by a specific entity.
Unit Market Share: The units sold by a particular company as a percentage of
total market sales, measured in the same units.
Unit Market Share (%) =
Unit Sales (#)
Total Market Unit Sales (#)
Chapter 2 Share of Hearts, Minds, and Markets
23
This formula, of course, can be rearranged to derive either unit sales or total market
unit sales from the other two variables, as illustrated in the following:
Unit Sales (#) = Unit Market Share (%) * Total Market Unit Sales (#)
Total Market Unit Sales (#) =
Unit Sales (#)
Unit Market Share (%)
Revenue Market Share: Revenue market share differs from unit market share in
that it reflects the prices at which goods are sold. In fact, a relatively simple way
to calculate relative price is to divide revenue market share by unit market share
(see Section 7.1).
Revenue Market Share (%) =
Sales Revenue ($)
Total Market Sales Revenue ($)
As with the unit market share, this equation for revenue market share can be rearranged
to calculate either sales revenue or total market sales revenue from the other two variables.
Data Sources, Complications, and Cautions
Market definition is never a trivial exercise: If a firm defines its market too broadly,
it may dilute its focus. If it does so too narrowly, it will miss opportunities and allow
threats to emerge unseen. To avoid these pitfalls, as a first step in calculating market share, managers are advised to define the served market in terms of unit sales or
revenues for a specific list of competitors, products, sales channels, geographic areas,
customers, and time periods. They might posit, for example, that “Among grocery
stores, we are the revenue market share leader in sales of frozen Italian food entrées in
the Northeastern U.S.”
Data parameters must be carefully defined: Although market share is likely the single
most used marketing metric, there is no generally acknowledged best method for calculating it. This is unfortunate, as different methods may yield not only different computations of market share at a given moment, but also widely divergent trends over time.
The reasons for these disparities include variations in the lenses through which share is
viewed (units versus dollars), where in the channel the measurements are taken (shipments from manufacturers versus consumer purchases), market definition (scope of the
competitive universe), and measurement error. In the situation analysis that underlies
strategic decisions, managers must be able to understand and explain these variations.
Competitive dynamics in the automobile industry, and at General Motors in particular,
illustrate the complexities involved in quantifying market share:
24
MARKETING METRICS
“With market share sliding in the first two months of the year, from 27.2% to 24.9%—
the lowest level since a two-month strike shut the company down in 1998—GM as a
whole expects a net loss of $846 million the first quarter.”2
Reviewing this statement, drawn from Business Week in 2005, a marketing manager
might immediately pose a number of questions:
■
Do these figures represent unit (autos sold) or revenue (dollar) market shares?
■
Does this trend hold for both unit and revenue market shares at GM?
■
Was revenue market share calculated before or after rebates and discounts?
■
■
Do the underlying sales data reflect factory shipments, which relate directly to
the manufacturer’s current income statement, or sales to consumers, which are
buffered by dealer inventories?
Does the decline in market share translate to an equivalent percentage decrease
in sales, or has the total market size changed?
Managers must determine whether a stated market share is based on shipment data,
channel shipments, retail sales, customer surveys, or some other source. On occasion,
share figures may represent combinations of data (a firm’s actual shipments, for example, set against survey estimates of competitors’ sales). If necessary, managers must also
adjust for differences in channels.
The time period measured will affect the signal-to-noise ratio: In analyzing shortterm market dynamics, such as the effects of a promotion or a recent price change,
managers may find it useful to measure market share over a brief period of time. Shortterm data, however, generally carry a low signal-to-noise ratio. By contrast, data covering a longer time span will be more stable but may obscure important, recent changes in
the market. Applied more broadly, this principle also holds in aggregating geographic
areas, channel types, or customers. When choosing markets and time periods for analysis, managers must optimize for the type of signal that is most important.
Potential bias in reported shares: One way to find data for market sizing is through
surveys of customer usage (see Section 2.7). In interpreting these data, however, managers must bear in mind that shares based on reported (versus recorded) sales tend to
be biased toward well-known brands.
Related Metrics and Concepts
Served Market: That portion of the total market for which the firm competes. This may
exclude geographic regions or product types. In the airline industry, for example, as of
mid-2015, Ryanair did not fly to the United States. Consequently, the U.S. would not be
considered part of its served market.
Chapter 2 Share of Hearts, Minds, and Markets
25
2.2 Relative Market Share and Market Concentration
Relative market share indexes a firm’s or a brand’s market share against that of its
leading competitor.
Brand’s Market Share (%)
Relative Market Share (I) =
Largest Competitor’s Market Share (%)
Market concentration, a related metric, measures the degree to which a comparatively
small number of firms accounts for a large proportion of the market.
These metrics are useful in comparing a firm’s or a brand’s relative position across
different markets and in evaluating the type and degree of competition in those
markets.
Purpose: To assess a firm’s or a brand’s success and its position
in the market.
A firm with a market share of 25% would be a powerful leader in many markets but
a distant “number two” in others. Relative market share offers a way to benchmark a
firm’s or a brand’s share against that of its largest competitor, enabling managers to
compare relative market positions across different product markets. Relative market
share gains some of its significance from studies—albeit controversial ones—suggesting
that major players in a market tend to be more profitable than their competitors. This
metric was further popularized by the Boston Consulting Group in its famous matrix of
relative share and market growth (see Figure 2.1).
High
Star
Question Mark or
Problem Child
Cash Cow
Dog
Market
Growth
Low
High
Relative Market Share
Figure 2.1
26
MARKETING METRICS
The BCG Matrix
Low
In the BCG matrix, one axis represents relative market share—a surrogate for competitive strength. The other represents market growth—a surrogate for potential.
Along each dimension, products are classified as high or low, placing them in one of
four quadrants. In the traditional interpretation of this matrix, products with high
relative market shares in growing markets are deemed stars, suggesting that they
should be supported with vigorous investment. The cash for that investment may be
generated by cash cows, products with high relative shares in low-growth markets.
Problem child products may have potential for future growth but hold weak competitive positions. Finally, dogs have neither strong competitive position nor growth
potential.
Construction
Relative Market Share (I) =
Brand’s Market Share (%)
Largest Competitor’s Market Share (%)
Relative market share can also be calculated by dividing brand sales (#,$) by largest
competitor’s sales (#,$) because the common factor of total market sales (or revenue)
cancels out.
EXAMPLE:
The market for small urban cars consists of five players (see Table 2.1).
Table 2.1 Market for Small Urban Cars
Units Sold (Thousands)
Revenue (Thousands)
Zipper
25
€375,000
Twister
10
€200,000
A-One
7.5
€187,500
Bowlz
5
€125,000
Chien
2.5
€50,000
50.0
€937,500
Market Total
In the market for small urban cars, managers at A-One want to know their firm’s market share
relative to its largest competitor. They can calculate this on the basis of revenues or unit sales.
In unit terms, A-One sells 7,500 cars per year. Zipper, the market leader, sells 25,000.
A-One’s relative market share in unit terms is thus 7,500/25,000 or 0.30. We arrive at the same
Chapter 2 Share of Hearts, Minds, and Markets
27
number if we first calculate A-One’s share (7,500/50,000 = .15) and Zipper’s share (25,000/50,000 =
.50) and then divide A-One’s share by Zipper’s share (.15/.50 = .30).
In revenue terms, A-One generates €187.5 million in car sales each year. Zipper, the market leader,
generates €375 million. A-One’s relative market share in revenue terms is thus €187.5m/€375m, or
0.5. Due to its comparatively high average price per car, A-One’s relative market share is greater in
revenue than in unit terms.
Related Metrics and Concepts
Market Concentration: The degree to which a relatively small number of firms
accounts for a large proportion of the market. This is also known as the concentration
ratio. It is usually calculated for the largest three or four firms in a market.3
Three (Four) Firm Concentration Ratio: A metric which is the total (sum) of the market shares held by the leading three (four) competitors in a market.
EXAMPLE: In the small urban car market, the three firm concentration ratio is comprised of
the market shares of the top three competitors—Zipper, Twister, and A-One (see Table 2.2).
Table 2.2 Market Share—Small Urban Cars
Units Sold
(Thousands)
Unit Share
Revenue
(Thousands)
Revenue Share
Zipper
25.0
50%
€375,000
40.0%
Twister
10
20%
€200,000
21.3%
A-One
7.5
15%
€187,500
20.0%
Bowlz
5.0
10%
€125,000
13.3%
Chien
2.5
5%
€50,000
5.3%
50.0
100%
€937,500
100%
Market Total
In unit terms, the three firm concentration ratio is 50% + 20% + 15% = 85%.
In revenue terms, it is 40% + 21.3% + 20% = 81.3%.
28
MARKETING METRICS
Herfindahl Index: A market concentration metric derived by adding the squares
of the individual market shares of all the players in a market. As a sum of
squares, this index tends to rise in markets dominated by large players.
EXAMPLE: The Herfindahl Index dramatically highlights market concentration in the small
urban car market (see Table 2.3).
Table 2.3 Calculation of the Herfindahl Index for Small Urban Cars
Units Sold
(Thousands)
Unit
Share
Squared
Share
Revenue
(Thousands)
Revenue
Share
Squared
Share
Zipper
25.0
50%
0.25
€375,000
40%
0.16
Twister
10
20%
0.04
€200,000
21%
0.0455
A-One
7.5
15%
0.0225
€187,500
20%
0.04
Bowlz
5.0
10%
0.01
€125,000
13%
0.0178
Chien
2.5
5%
0.0025
€50,000
5%
0.0028
Market Total 50.0
100%
0.325
€937,500
100%
0.2661
On a unit basis, the Herfindahl Index is equal to the square of the unit market share of Zipper (50%
^ 2 = 0.25), plus that of Twister (20% ^ 2 = 0.04), plus those of A-One, Bowlz, and Chien = 0.325.
On a revenue basis, the Herfindahl Index comprises the square of the revenue market share of
Zipper (40% ^ 2 = 0.16), plus those of all its competitors = 0.2661.
As demonstrated by the Herfindahl Index, the market for small urban cars is slightly more concentrated in unit terms than in revenue terms. The reason for this is straightforward: Higher-priced
cars in this market sell fewer units.
Note: For a given number of competitors, the Herfindahl Index would be lowest if shares were
equally distributed. In a five-firm industry, for example, equally distributed shares would yield a
Herfindahl Index of 5 * (20% ^ 2) = 0.2.
Data Sources, Complications, and Cautions
As ever, appropriate market definition and the use of comparable figures are vital prerequisites to developing meaningful results.
Chapter 2 Share of Hearts, Minds, and Markets
29
Related Metrics and Concepts
Market Share Rank: The ordinal position of a brand in its market, when competitors
are arranged by size, with 1 being the largest.
Share of Category: This metric is derived in the same manner as market share, but
is used to denote a share of market within a certain retailer or class of retailers (for
example, mass merchandisers).
2.3 Brand Development Index and Category
Development Index
The brand development index (BDI) quantifies how well a brand is performing
within a specific group of customers, compared with its average performance among
all consumers.
Brand Development Index (I) =
Brand Sales to Group (#,$)/Households (#)
in the Group
Total Brand Sales (#,$)/Total Households (#)
The category development index (CDI) measures the sales performance of a category
of goods or services within a specific group, compared with its average performance
among all consumers.
Category Sales to Group (#,$)/Households in Group (#)
Category Development Index (I) =
Total Category Sales (#,$)/Total Households (#)
The brand and category development indexes are useful for understanding specific customer segments relative to the market as a whole. Although defined here
with respect to households, these indexes could also be calculated for customers,
accounts, businesses, or other entities.
Purpose: To understand the relative performance of a brand
or category within specified customer groups.
The brand and category development indexes help identify strong and weak segments
(usually, demographic or geographic) for particular brands or categories of goods and
services. For example, by monitoring the CDI (category development index), marketers
might determine that Midwesterners buy twice as many country-western music CDs
per capita as Americans in general, while consumers living on the East Coast buy less
30
MARKETING METRICS
than the national average. This would be useful information for targeting the launch
campaign for a new country-western performer. Conversely, if managers found that a
particular product had a low brand development index in a segment that carried a high
CDI for its category, they might ask why that brand suffered relatively poor performance in such a promising segment.
Construction
Brand Development Index—BDI (I): An index of how well a brand performs within a
given market group, relative to its performance in the market as a whole.
Brand Development Index—BDI (I) =
Brand Sales to Group (#,$)/ Households in Group (#)
Total Brand Sales (#,$)/ Total Households (#)
The BDI (brand development index) is a measure of brand sales per person or per
household within a specified demographic group or geography, compared with its
average sales per person or household in the market as a whole. To illustrate its use:
One might hypothesize that sales per capita of Ben & Jerry’s brand ice cream would
be greater in the brand’s home state, Vermont, than in the rest of the country. By
calculating Ben & Jerry’s BDI for Vermont, marketers could test this hypothesis
quantitatively.
EXAMPLE: Oaties is a minor brand of breakfast cereal. Among households without children,
its sales run one packet per week per 100 households. In the general population, Oaties’ sales run
one packet per week per 80 households. This translates to 1/100 of a packet per household in the
childless segment, versus 1/80 of a packet in the general populace.
BDI =
=
Brand Sales/Households
Total Brand Sales/Households
1/100
1/80
= 0.8
Oaties performs slightly less well in the childless segment than in the market as a whole.
Category Development Index—CDI: An index of how well a category performs
within a given market segment, relative to its performance in the market as a
whole.
Chapter 2 Share of Hearts, Minds, and Markets
31
Category Development Index (I) =
Category Sales to Group (#,$)/Households in Group (#)
Total Category Sales (#,$)/Total Households (#)
Similar in concept to the BDI, the category development index demonstrates where
a category shows strength or weakness relative to its overall performance. By way of
example, Boston enjoys high per-capita consumption of ice cream. Bavaria and Ireland
both show higher per-capita consumption of beer than Iran.
Data Sources and Complications
In calculating BDI or CDI, a precise definition of the segment under study is vital.
Segments are often bounded geographically, but they can be defined in any way for
which data can be obtained.
Related Metrics and Concepts
The term category development index has also been applied to retail organizations.
In this application, it measures the extent to which a retailer emphasizes one category
versus others.
Category Development Index (I) =
Retailer’s Share of Category Sales (%)
Retailer’s Total Share of Market (%)
This use of the term is similar to the category performance ratio (see Section 6.6).
2.4 Penetration
Penetration is a measure of brand or category popularity. It is defined as the number
of people who buy a specific brand or a category of goods at least once in a given
period, divided by the size of the relevant market population.
Market Penetration (%) =
Customers Who Have Purchased a Product in the Category (#)
Brand Penetration (%) =
Total Population (#)
Customers Who Have Purchased the Brand (#)
Penetration Share (%) =
32
MARKETING METRICS
Total Population (#)
Brand Penetration (%)
Market Penetration (%)
Penetration Share (%) =
Customers Who Have Purchased the Brand (#)
Customers Who Have Purchased a Product in the Category (#)
Often, managers must decide whether to seek sales growth by acquiring existing category users from their competitors or by expanding the total population of category
users, attracting new customers to the market. Penetration metrics help indicate
which of these strategies would be most appropriate and help managers to monitor
their success. These equations might also be calculated for usage instead of purchase.
Construction
Penetration: The proportion of people (households) in the relevant population
who bought (at least once in the period) a specific brand or a category of goods.
Market Penetration (%) =
Customers Who Have Purchased a Product in the Category (#)
Brand Penetration (%) =
Total Population (#)
Customers Who Have Purchased the Brand (#)
Total Population (#)
Whereas market share focuses on the sales of a product (either units or dollars), penetration focuses on the number of buyers. As such, it is a measure of popularity.
EXAMPLE: In the previous month, 500 of the 10,000 households in the panel purchased Big
Bomb Flea Flogger, and 750 bought any kind of flea flogger, including Big Bomb,
Big Bomb’s Brand Penetration = 500/10,000 = 5%
Flea Flogger Category Penetration = 750/10,000 = 7.5%
Big Bomb’s Penetration Share = 500/750 = 66.67%.
It is not surprising that in a given month, only 7.5% of the panel bought flea spray. The amount of
penetration to expect is a combination of how often the product is bought and the length of the
time period. The longer the time period, the higher will be the calculated penetration.
Penetration share can be thought of as the share of category households that bought the
brand. Thus a brand’s penetration share can never be greater than one. When households buy multiple brands, however, penetration shares will sum to more than one.
Chapter 2 Share of Hearts, Minds, and Markets
33
Penetration Share (%) =
Brand Penetration (%)
Market Penetration (%)
Decomposing Market Share
Market share can always be calculated as the product of three components: penetration
share, share of requirements, and usage index.
Market Share (%) = Penetration Share (%) * Share of Requirements (%)
* Usage Index (I)
This decomposition is useful in that it identifies three ways to improve market share:
sell to more people, achieve a higher share of your customer’s purchases, or get your
customers to use more of the category. Although this is true by definition, in practice
the three components rarely move independently.
This decomposition works for both revenue and unit share depending on whether
share of requirements is calculated using units or dollars.
There are four variables in this decomposition, and (like any equation) it can be used to
find the fourth if the other three are known. In subsequent sections we will give equations for each component in terms of market share and the remaining two components.
Penetration Share (%) =
Market Share (%)
[Usage Index (I) * Share of Requirements (%)]
EXAMPLE: Eat Wheats brand cereal has a market share in Urbanopolis of 6%, a usage index of
0.75 and a share of requirements of 40%. From these data, we can calculate the penetration share for
Eat Wheats brand cereal in Urbanopolis:
Penetration Share =
=
Market Share
Usage Index * Share of Requirements
6%
0.75 * 40%
=
6%
.30
= 20%
Data Sources, Complications, and Cautions
The time period over which a firm measures penetration will have a significant impact
on the result. For example, even among the most popular detergent brands, many are
not purchased weekly. As the time period used to define penetration becomes shorter,
34
MARKETING METRICS
brand penetration declines. In contrast, although penetration share will be more volatile for shorter periods it will not necessarily be lower.
Related Metrics and Concepts
Total Number of Active Customers: The customers (accounts) who purchased at
least once in a given time period. When assessed at a brand level, this is equivalent to
brand penetration. This term is often used in shorthand form—total number of customers—though this would not be appropriate when a distinction must be made for
ex-customers. This is discussed in more detail in Section 5.1 (customers of a specified
recency).
Accepters: Customers who are disposed to accept a given product and its benefits: the
opposite of rejecters.
Ever-tried: The percentage of a population that has tried a given brand at any time.
(See Section 4.1 for more on trial.)
2.5 Share of Requirements
Share of requirements, also known as share of wallet, is calculated solely among
buyers of a specific brand. Within this group, it represents the percentage of purchases within the relevant category, accounted for by the brand in question.
Unit Share of Requirements (%) =
Brand Purchases (#)
Total Category Purchases by Brand Buyers (#)
Revenue Share of Requirements (%) =
Brand Purchases ($)
Total Category Purchases by Brand Buyers ($)
Many marketers view share of requirements as a measure of loyalty. This metric can
guide a firm’s decisions on whether to allocate resources toward efforts to expand a
category, to take customers from competitors, or to increase share of requirements
among its established customers. Share of requirements is, in essence, the market
share a brand enjoys within a market narrowly defined as the people who have purchased that brand.
Chapter 2 Share of Hearts, Minds, and Markets
35
Purpose: To understand the source of market share in terms of breadth and
depth of consumer franchise, as well as the extent of relative category usage
(heavy users/larger customers versus light users/smaller customers).
Construction
Share of Requirements: A given brand’s share of purchases in its category,
measured solely among customers who have purchased that brand. Also known
as share of wallet.
When calculating share of requirements, marketers may use either dollars or units.
Unit Share of Requirements (%) =
Brand Purchases (#)
Total Category Purchases by Brand Buyers (#)
Revenue Share of Requirements (%) =
Brand Purchases ($)
Total Category Purchases by Brand Buyers ($)
The best way to think about share of requirements is as the market share enjoyed by a
product among the customers who buy it.
EXAMPLE: In a given month, the unit purchases of AloeHa brand sunscreen ran 1,000,000
bottles. Among the households that bought AloeHa, total purchases of sunscreen came to 2,000,000
bottles.
Share of Requirements =
=
AloeHa Purchases
Category Purchases by AloeHa Customers
1,000,000
2,000,000
= 50%
Share of requirements is also useful in analyzing overall market share. As previously
noted, it is part of an important formulation of market share.
Market Share = Penetration Share * Share of Requirements * Usage Index
Share of requirements can thus be calculated indirectly by decomposing market share.
36
MARKETING METRICS
Share of Requirements (%) =
Market Share (%)
[Penetration Share (%) * Usage Index (I)]
EXAMPLE: Eat Wheats brand cereal has a market share in Urbanopolis of 8%. The usage index
for Eat Wheats in Urbanopolis is 1. The brand’s penetration share in Urbanopolis is 20%. On this
basis, we can calculate Eat Wheats’ share of requirements in Urbanopolis:
Share of Requirements =
=
Market Share
(Usage Index * Penetration Share)
8%
(1 * 20%)
=
8%
20%
= 40%
Note that in this example, market share and usage index must both be defined in the same terms
(units or revenue). Depending on the definition of these two metrics, the calculated share of
requirements will be either unit share of requirements (%) or revenue share of requirements (%).
Data Sources, Complications, and Cautions
Double Jeopardy: As mentioned earlier, the three components of market share do not
move independently in practice. In an empirical observation labeled as “double jeopardy,” brands with lower market share almost invariably have lower penetration share
and lower share of requirements. The name double jeopardy captures the notion that
low share brands are punished twice for their lower share. Not only do fewer households
buy them, those buying households buy less of them (lower share of requirements).
Double jeopardy is a very real and pervasive phenomenon. One implication is that
improvements in market share will lead to improvements in both penetration and
share of requirements. It is also true that brands appear to vary more with respect to
penetration than share of requirements—which has led others to argue that it is better
to try to increase penetration (sell to more households) than to increase the loyalty of
the brand’s current customers. We do not endorse that argument.
One explanation for double jeopardy is that low share brands do not get broad distribution. Thus it is more difficult for the few customers who prefer them to find and buy
them, and share of requirements suffers. In this way of thinking, it is share that leads to
both penetration and share of requirements.
Chapter 2 Share of Hearts, Minds, and Markets
37
There is also a purely statistical explanation for double jeopardy that we will describe at
the end of Section 2.6.
Related Metrics and Concepts
Sole Usage Percentage: The proportion of a brand’s customers who bought only that
brand’s product and did not buy from competitors. Sole users are a combination of
extremely loyal customers, customers with limited access to brands, and customers who
just happened to buy one unit during the period. Sole users are also known as purely
loyals. Among sole users, a brand’s share of requirement is, by definition, 100%.
Sole Usage (%) =
Customers Who Buy Only the Brand in Question (#)
Total Brand Customers (#)
Number of Brands Purchased: During a given period, some customers may buy only
a single brand within a category, whereas others buy two or more. In evaluating loyalty to
a given brand, marketers sometimes compare the average number of brands purchased
by brand customers to the average number purchased by all customers in the category.
EXAMPLE: Among 10 customers for cat food, 7 bought the Arda brand, 5 bought Bella, and
3 bought Constanza. One does not need to know the pattern of overlap across these customers to
calculate the average number of brands per cat food customer. It is 15 brand purchases (7 + 5 + 3)
per ten customers, yielding an average of 1.5 brands per customer.
A Bella brand manager noted that of his firm’s five customers, 3 bought only Bella, whereas two
bought both Arda and Bella. None of Bella’s customers bought Constanza. Thus, the five Bella
customers made seven brand purchases (1 + 1 + 1 + 2 + 2), yielding an average of 1.4 (that is,
7/5) brands per Bella customer. Compared to the average category purchaser, who buys 1.5 brands,
Bella buyers are slightly more loyal.
Repeat Rate: The percentage of brand customers in a given period who are also
brand customers in the subsequent period.
Repurchase Rate: The percentage of customers for a brand who repurchase that
brand on their next purchase occasion.
Confusion abounds in this area. In these definitions, we have tried to distinguish a
metric based on calendar time (repeat rate) from one based on “customer time” (repurchase rate). In Chapter 5, “Customer Profitability,” we will describe a related metric,
38
MARKETING METRICS
retention, which is used in contractual situations in which the first non-renewal (nonpurchase) signals the end of a customer relationship. Although we suggest that the term
retention be applied only in contractual situations, you will often see repeat rates and
repurchase rates referred to as “retention rates.” Due to a lack of consensus on the use
of these terms, marketers are advised not to rely on the names of these metrics as perfect indicators of how they are calculated.
As with penetration, the interpretation of repeat rate depends on the time period
covered. The shorter the time period, the lower will be the repeat rate. One minus the
repeat rate is sometimes called the turnover rate.
2.6 Usage Index
The usage index is a relative measure that indicates how heavily the customers of a
given brand use the product category.
Usage Index (I) =
Average Total Purchases in Category by Brand Customers (#,$)
Average Total Purchases in Category by All Customers for
That Category (#,$)
or
Usage Index (I) =
Market Share (%)
[Penetration Share (%) * Share of Requirements (%)]
The usage index, also called the weight index, yields insight into the source of volume and the nature of a brand’s customer base.
Purpose: To define and measure whether a firm’s consumers
are “heavy users.”
The usage index answers the question, “How heavily do our customers use the category
of our product?”
Construction
Usage Index: The ratio of the average category usage for the customers of a
brand to the average category usage of all customers.
The usage index can be calculated on the basis of unit or dollars.
Chapter 2 Share of Hearts, Minds, and Markets
39
Average Total Purchases in Category by Brand Customers (#,$)
Usage Index (I) =
Average Total Purchases in Category by All Customers
for That Category (#,$)
EXAMPLE: Over a period of one year, the average shampoo purchases by households using
Shower Fun brand shampoo was six 15-oz bottles. During the same period, average shampoo consumption by households using any brand of shampoo was four 15-oz bottles.
The usage index for households buying Shower Fun is therefore 6/4, or 1.5. Customers of Shower
Fun brand shampoo bought more shampoo than average.
As previously noted, market share can be calculated as the product of three components: penetration share, share of requirements, and usage index (see Section 2.4).
Consequently, we can calculate a brand’s usage index if we know its market share, penetration share, and share of requirements, as follows:
Usage Index (I) =
Market Share (%)
Penetration Share (%) * Share of Requirements (%)
This equation works for market shares defined in either unit or dollar as long as the
construction of share of requirements and usage index matches. Comparing a brand’s
dollar usage index to its unit usage index, marketers can determine whether the brand
purchasers pay more (less) per category unit.
Data Sources, Complications, and Cautions
The usage index does not indicate how heavily customers use a specific brand, only how
heavily they use the category. A brand can have a high usage index, for example, meaning that its customers are heavy category users, even if those customers use the brand in
question to meet only a small share of their needs.
Understanding Double Jeopardy and Usage Index
In the previous section we described an empirical generalization called double jeopardy. Brands with lower share are punished twice in that they sell to fewer people (lower
penetration) who buy them less (lower share of requirements). Before jumping to conclusions about the marketing strategy implications of double jeopardy, however, it is
important to understand the statistical explanation for double jeopardy.
40
MARKETING METRICS
Double jeopardy is a natural consequence of two things:
■
■
How the component metrics are defined (with customers of a brand being
defined as all households that bought the brand during the period).
Households often buy multiple brands within a given period for inexplicable
reasons.
To illustrate, consider a market with three brands (A, B, C) serving ten households with
unit sales for the period given in Table 2.4.
Table 2.4 Example Household Brand Purchase Counts
Household
A
B
C
Total
1
10
0
0
10
2
10
0
0
10
3
10
0
0
10
4
10
0
0
10
5
10
0
0
10
6
10
0
0
10
7
0
10
0
10
8
0
10
0
10
9
0
10
0
10
10
0
0
10
10
The category summary metrics paint a clear picture of this ten household market. All
households were loyal to a single brand and all households bought the same number of
units. The 60/30/10 share split exactly matches the brands’ penetrations (PENs) of 60%,
30%, and 10% respectively (see Table 2.5).
Table 2.5
Example Summary Metrics
Summary Metric
A
B
C
Share of Market
0.6
0.3
0.1
Penetration
0.6
0.3
0.1
Share of Requirements
1.0
1.0
1.0
Usage Index
1.0
1.0
1.0
Chapter 2 Share of Hearts, Minds, and Markets
41
But now let us add an 11th household that is a heavy, but indiscriminate, buyer purchasing 12 units of A, 6 units of B, and 2 units of C (see Table 2.6).
Table 2.6 Example Household Brand Purchase Counts After Adding New Customer
Household
A
B
C
Total
1
10
0
0
10
2
10
0
0
10
3
10
0
0
10
4
10
0
0
10
5
10
0
0
10
6
10
0
0
10
7
0
10
0
10
8
0
10
0
10
9
0
10
0
10
10
0
0
10
10
11
12
6
2
20
We contend that the presence of this multiple-brand buyer doesn’t really change our
view of the relative loyalties enjoyed by the three brands. Brand C got a 10% share of
the ten loyal households and also a 10% share of not-so-loyal household 11. Brand C
performed exactly as one might expect.
But look at the new category summary statistics (see Table 2.7):
Table 2.7
Example Summary Metrics After Adding New Customer
Summary Metric
A
B
C
Share of Market
0.600
0.300
0.100
Penetration
0.636
0.364
0.182
Usage Index
1.048
1.146
1.375
Share of Requirements
0.900
0.720
0.400
42
MARKETING METRICS
The penetration metrics have all gone up a bit and now sum to something greater than
one because household 11 is counted as a customer of each of the three brands. A consequence of this double counting and the fact that household 11 was a heavy category
user is that all three usage indices are greater than one. In our experience this is often
the case. Heavy using households tend to buy several brands (which makes sense) and
we end up with usage indices that are all greater than one.
It also makes sense that the presence of household 11 causes share of requirements to
decrease for all three brands. But the important observation is that the decrease in share
of requirements is much greater for the small share brand simply because it is the small
share brand and heavy user household 11 has a bigger influence on its summary statistics. (Household 11 is 1/2 of brand C’s customers but only 1/7 of brand A’s.) The result
is double jeopardy caused by the presence of household 11 and the fact that household
11 gets counted as a customer of all three brands. Although this is appropriate in one
sense (household 11 is a customer for all three brands) and allows our market share
decomposition to hold, it affects the behavior of share of requirements and usage index
metrics in a way that makes them challenging to interpret. Certainly for us, share of
requirements is a less than perfect measure of the concept of loyalty.
Related Metrics and Concepts
See also the discussion of brand development index (BDI) and category development
index (CDI) in Section 2.3.
2.7 Awareness, Attitudes, and Usage (AAU): Metrics
of the Hierarchy of Effects
Studies of awareness, attitudes, and usage (AAU) enable marketers to quantify levels
and trends in customer knowledge, perceptions, beliefs, intentions, and behaviors. In
some companies, the results of these studies are called “tracking” data because they
are used to track long-term changes in customer awareness, attitudes, and behaviors.
AAU studies are most useful when their results are set against a clear comparator. This benchmark may comprise the data from prior periods, different markets,
or competitors.
Purpose: To track trends in customer attitudes and behaviors.
Awareness, attitudes, and usage (AAU) metrics relate closely to what has been called
the Hierarchy of Effects, an assumption that customers progress through sequential
stages from lack of awareness, through initial purchase of a product, to brand loyalty
Chapter 2 Share of Hearts, Minds, and Markets
43
(see Figure 2.2). AAU metrics are generally designed to track these stages of knowledge,
beliefs, and behaviors. AAU studies also may track “who” uses a brand or product—in
which customers are defined by category usage (heavy/ light), geography, demographics, psychographics, media usage, and whether they purchase other products.
Awareness
Customers must first become aware of a product, then . . .
Attitudes
They develop attitudes and beliefs about that product, and finally . . .
Usage
Figure 2.2
Customers purchase and experience the product.
Awareness, Attitudes, and Usage: Hierarchy of Effects
Information about attitudes and beliefs offers insight into the question of why specific
users do, or do not, favor certain brands. Typically, marketers conduct surveys of large
samples of households or business customers to gather these data.
Construction
Awareness, attitudes, and usage studies feature a range of questions that aim to shed
light on customers’ relationships with a product or brand (see Table 2.8). For example,
who are the acceptors and rejecters of the product? How do customers respond to a
replay of advertising content?
Table 2.8 Awareness, Attitudes, and Usage: Typical Questions
Type
Measures
Typical Questions
Awareness
Awareness and
Knowledge
Have you heard of Brand X?
Attitudes
Beliefs and
Intentions
Is Brand X for me?
On a scale of 1 to 5, is Brand X for young people?
What are the strengths and weaknesses of each brand?
Usage
Purchase Habits Did you use Brand X this week?
and Loyalty
What brand did you last buy?
44
MARKETING METRICS
What brand comes to mind when you think “luxury car?”
Marketers use answers to these questions to construct a number of metrics. Among
these, certain “summary metrics” are considered important indicators of performance.
In many studies, for example, customers’ “willingness to recommend” and “intention to
purchase” a brand are assigned high priority. Underlying these data, various diagnostic
metrics help marketers understand why consumers may be willing—or unwilling—to
recommend or purchase that brand. Consumers may not have been aware of the brand,
for example. Alternatively, they may have been aware of it but did not subscribe to one
of its key benefit claims.
Awareness and Knowledge
Marketers evaluate various levels of awareness, depending on whether the consumer
in a given study is prompted by a product’s category, brand, advertising, or usage
situation.
Awareness: The percentage of potential customers or consumers who recognize—
or name—a given brand. Marketers may research brand recognition on an “aided”
or “prompted” level, posing such questions as, “Have you heard of Mercedes?”
Alternatively, they may measure “unaided” or “unprompted” awareness, posing
such questions as, “Which makes of automobiles come to mind?”
Top of Mind: The first brand that comes to mind when a customer is asked an
unprompted question about a category. The percentage of customers for whom a
given brand is top of mind can be measured.
Ad Awareness: The percentage of target consumers or accounts who demonstrate awareness (aided or unaided) of a brand’s advertising. This metric can be
campaign- or media-specific, or it can cover all advertising.
Brand/Product Knowledge: The percentage of surveyed customers who demonstrate specific knowledge or beliefs about a brand or product.
Attitudes
Measures of attitude concern consumer response to a brand or product. Attitude is a
combination of what consumers believe and how strongly they feel about it. Although a
detailed exploration of attitudinal research is beyond the scope of this book, the following summarizes certain key metrics in this field.
Attitudes/Liking/Image: A rating assigned by consumers—often on a scale of
1–5 or 1–7—when survey respondents are asked their level of agreement with
such propositions as, “This is a brand for people like me,” or “This is a brand for
young people.” A metric based on such survey data can also be called relevance to
customer.
Chapter 2 Share of Hearts, Minds, and Markets
45
Perceived Value for Money: A rating assigned by consumers—often on a scale of
1–5 or 1–7—when survey respondents are asked their level of agreement with such
propositions as, “This brand usually represents a good value for the money.”
Perceived Quality/Esteem: A consumer rating—often on a scale of 1–5 or 1–7—
of a given brand’s product when compared with others in its category or market.
Relative Perceived Quality: A consumer rating (often from 1–5 or 1–7) of brand
product compared to others in the category/market.
Intentions: A measure of customers’ stated willingness to behave in a certain way.
Information on this subject is gathered through such survey questions as, “Would
you be willing to switch brands if your favorite was not available?”
Purchase Intentions: A specific measure or rating of consumers’ stated purchase
intentions. Information on this subject is gathered through survey respondents’
reactions to such propositions as, “It is very likely that I will purchase this
product.”
Usage
Measures of usage concern such market dynamics as purchase frequency and units
per purchase. They highlight not only what was purchased, but also when and
where it was purchased. In studying usage, marketers also seek to determine how
many people have tried a brand. Of those, they further seek to determine how many
have “rejected” the brand, and how many have “adopted” it into their regular portfolio of brands.
Usage: A measure of customers’ self-reported behavior.
In measuring usage, marketers pose such questions as the following: What brand of
toothpaste did you last purchase? How many times in the past year have you purchased
toothpaste? How many tubes of toothpaste do you currently have in your home? Do
you have any Crest toothpaste in your home at the current time?
In the aggregate, AAU metrics concern a vast range of information that can be tailored
to specific companies and markets. They provide managers with insight into customers’
overall relationships with a given brand or product.
46
MARKETING METRICS
Data Sources, Complications, and Cautions
Sources of AAU data include
■
■
Warranty cards and registrations, often using prizes and random drawings to
encourage participation.
Regularly administered surveys, conducted by organizations that interview
consumers via telephone, mail, Web, or other technologies, such as hand-held
scanners.
Even with the best methodologies, however, variations observed in tracking data from
one period to the next are not always reliable. Managers must rely on their experience to distinguish seasonality effects and “noise” (random movement) from “signal”
(actual trends and patterns). Certain techniques in data collection and review can also
help managers make this distinction.
1. Adjust for periodic changes in how questions are framed or administered.
Surveys can be conducted via mail or telephone, for example, among paid or
unpaid respondents. Different data-gathering techniques may require adjustment in the norms used to evaluate a “good” or “bad” response. If sudden
changes appear in the data from one period to the next, marketers are advised
to determine whether methodological shifts might play a role in this result.
2. Try to separate customer from non-customer responses; they may be very different. Causal links among awareness, attitudes, and usage are rarely clear-cut.
Though the hierarchy of effects is often viewed as a one-way street, on which
awareness leads to attitudes, which in turn determine usage, the true causal
flow might also be reversed. When people own a brand, for example, they may
be predisposed to like it.
3. Triangulate customer survey data with sales revenue, shipments, or other data
related to business performance. Consumer attitudes, distributor and retail
sales, and company shipments may move in different directions. Analyzing
these patterns can be a challenge but can reveal much about category dynamics. For example, toy shipments to retailers often occur well in advance of the
advertising that drives consumer awareness and purchase intentions. These, in
turn, must be established before retail sales. Adding further complexity, in the
toy industry, the purchaser of a product might not be its ultimate consumer.
In evaluating AAU data, marketers must understand not only the drivers of
demand but also the logistics of purchase.
4. Separate leading from lagging indicators whenever possible. In the auto
industry, for example, individuals who have just purchased a new car show a
heightened sensitivity to advertisements for its make and model. Conventional
wisdom suggests that they’re looking for confirmation that they made a good
Chapter 2 Share of Hearts, Minds, and Markets
47
choice in a risky decision. By helping consumers justify their purchase at this
time, auto manufacturers can strengthen long-term satisfaction and willingness
to recommend.
Related Metrics and Concepts
Likeability: Because AAU considerations are so important to marketers, and because
there is no single “right” way to approach them, specialized and proprietary systems
have been developed. Of these, one of the best known is the Q scores rating of “likeability.” A Q Score is derived from a general survey of selected households, in which a large
panel of consumers share their feelings about brands, celebrities, and television shows.4
Q Scores rely upon responses reported by consumers. Consequently, although the
system used is sophisticated, it is dependent on consumers understanding and being
willing to reveal their preferences.
Segmentation by Geography, or Geo-clustering: Marketers can achieve insight into
consumer attitudes by separating their data into smaller, more homogeneous groups of
customers. One well-known example of this is Prizm. Prizm assigns U.S. households to
clusters based on ZIP Code,5 with the goal of creating small groups of similar households. The typical characteristics of each Prizm cluster are known, and these are used to
assign a name to each group. “Golden Ponds” consumers, for example, comprise elderly
singles and couples leading modest lifestyles in small towns. Rather than monitoring
AAU statistics for the population as a whole, firms often find it useful to track these
data by cluster.
2.8 Customer Satisfaction and Willingness
to Recommend
Customer satisfaction is generally based on survey data and expressed as a rating.
For example, see Figure 2.3.
Very
Dissatisfied
Somewhat
Dissatisfied
Neither Satisfied
nor Dissatisfied
Somewhat
Satisfied
Very
Satisfied
1
2
3
4
5
Figure 2.3 Ratings
48
MARKETING METRICS
Within organizations, customer satisfaction ratings can have powerful effects.
They focus employees on the importance of fulfilling customers’ expectations.
Furthermore, when these ratings dip, they warn of problems that can affect sales and
profitability.
A second important metric related to satisfaction is willingness to recommend.
When a customer is satisfied with a product, he or she might recommend it to
friends, relatives, and colleagues. This can be a powerful marketing advantage.
Purpose: Customer satisfaction provides a leading indicator of consumer
purchase intentions and loyalty.
Customer satisfaction data are among the most frequently collected indicators of market perceptions. Their principal use is twofold.
1. Within organizations, the collection, analysis, and dissemination of these data
send a message about the importance of tending to customers and ensuring
that they have a positive experience with the company’s goods and services.
2. Although sales or market share can indicate how well a firm is performing
currently, satisfaction is perhaps the best indicator of how likely it is that the
firm’s customers will make further purchases in the future. Much research
has focused on the relationship between customer satisfaction and retention.
Studies indicate that the ramifications of satisfaction are most strongly realized
at the extremes. On the scale in Figure 2.3, individuals who rate their satisfaction level as “5” are likely to become return customers and might even evangelize for the firm. Individuals who rate their satisfaction level as “1,” by contrast,
are unlikely to return. Further, they can hurt the firm by making negative
comments about it to prospective customers. Willingness to recommend is a
key metric relating to customer satisfaction.
Construction
Customer Satisfaction: The number of customers, or percentage of total customers,
whose reported experience with a firm, its products, or its services (ratings) exceeds
specified satisfaction goals.
Willingness to Recommend: The percentage of surveyed customers who indicate that
they would recommend a brand to friends.
These metrics quantify an important dynamic. When a brand has loyal customers, it
gains positive word-of-mouth marketing, which is both free and highly effective.
Chapter 2 Share of Hearts, Minds, and Markets
49
Customer satisfaction is measured at the individual level, but it is almost always
reported at an aggregate level. It can be, and often is, measured along various dimensions. A hotel, for example, might ask customers to rate their experience with its front
desk and check-in service, with the room, with the amenities in the room, with the restaurants, and so on. Additionally, in a holistic sense, the hotel might ask about overall
satisfaction “with your stay.”
Customer satisfaction is generally measured on a five-point scale (see Figure 2.4).
Very
Dissatisfied
Somewhat
Dissatisfied
Neither Satisfied
nor Dissatisfied
1
2
3
Somewhat
Satisfied
4
Very
Satisfied
5
Figure 2.4 A Typical Five-Point Scale
Satisfaction levels are usually reported as either “top box” or, more likely, “top
two boxes.” Marketers convert these expressions into single numbers that show the
percentage of respondents who checked either a “4” or a “5.” (This term is the same as
that commonly used in projections of trial volumes; see Section 4.1.)
EXAMPLE: The general manager of a hotel in Quebec institutes a new system of customer
satisfaction monitoring (see Figure 2.5). She leaves satisfaction surveys at checkout. As an incentive
to respond, all respondents are entered into a drawing for a pair of free airline tickets.
Score
Responses
(200 useable)
%
Very
Dissatisfied
Somewhat
Dissatisfied
Neither Satisfied Somewhat
nor Dissatisfied Satisfied
Very
Satisfied
1
2
3
4
5
3
7
40
100
50
1.5%
3.5%
20%
50%
25%
Figure 2.5
Hotel Customer Survey Response
The manager collects 220 responses, of which 20 are unclear or otherwise unusable. Among the
remaining 200, 3 people rate their overall experience at the hotel as very unsatisfactory, 7 deem
it somewhat unsatisfactory, and 40 respond that they are neither satisfied nor dissatisfied. Of the
remainder, 50 customers say they are very satisfied, while the rest are somewhat satisfied.
The top box, comprising customers who rate their experience a “5,” includes 50 people or, as a
percentage, 50/200 = 25%. The top two boxes comprise customers who are “somewhat” or “very”
50
MARKETING METRICS
satisfied, rating their experience a “4” or “5.” In this example, the “somewhat satisfied” population
must be calculated as the total usable response pool, less customers accounted for elsewhere; that
is, 200 – 3 – 7 – 40 – 50 = 100. The sum of the top two boxes is thus 50 + 100 = 150 customers, or
75% of the total.
Customer satisfaction data can also be collected on a 10-point scale. Regardless of the
scale used, the objective is to measure customers’ perceived satisfaction with their experience of a firm’s offerings. Marketers then aggregate these data into a percentage of
top-box responses.
In researching satisfaction, firms generally ask customers whether their product or
service has met or exceeded expectations. Thus, expectations are a key factor behind
satisfaction. When customers have high expectations and the reality falls short, they
will be disappointed and will likely rate their experience as less than satisfying. For
this reason, a luxury resort, for example, might receive a lower satisfaction rating than
a budget motel—even though its facilities and service would be deemed superior in
“absolute” terms.
Data Sources, Complications, and Cautions
Surveys constitute the most frequently used means of collecting satisfaction data. As a
result, a key risk of distortion in measures of satisfaction can be summarized in a single
question: Who responds to surveys?
“Response bias” is endemic in satisfaction data. Disappointed or angry customers
often welcome a means to vent their opinions. Contented customers often do not.
Consequently, although many customers might be happy with a product and feel no
need to complete a survey, the few who had a bad experience might be disproportionately represented among respondents. Most hotels, for example, place response cards in
their rooms, asking guests, “How was your stay?’ Only a small percentage of guests ever
bother to complete those cards. Not surprisingly, those who do respond probably had a
bad experience. For this reason, marketers can find it difficult to judge the true level of
customer satisfaction. By reviewing survey data over time, however, they may discover
important trends or changes. If complaints suddenly rise, for example, that may constitute early warning of a decline in quality or service. (See number of complaints in the
following section.)
Sample selection may distort satisfaction ratings in other ways as well. Because only
customers are surveyed for customer satisfaction, a firm’s ratings may rise artificially
as deeply dissatisfied customers take their business elsewhere. Also, some populations
may be more frank than others, or more prone to complain. These normative differences can affect perceived satisfaction levels. In analyzing satisfaction data, a firm might
Chapter 2 Share of Hearts, Minds, and Markets
51
interpret rating differences as a sign that one market is receiving better service than
another, when the true difference lies only in the standards that customers apply. To
correct for this issue, marketers are advised to review satisfaction measures over time
within the same market.
A final caution: Because many firms define customer satisfaction as “meeting or
exceeding expectations,” this metric may fall simply because expectations have risen.
Thus, in interpreting ratings data, managers may come to believe that the quality of
their offering has declined when that is not the case. Of course, the reverse is also true.
A firm might boost satisfaction by lowering expectations. In so doing, however, it might
suffer a decline in sales as its product or service comes to appear unattractive.
Related Metrics and Concepts
Trade Satisfaction: Founded upon the same principles as consumer satisfaction, trade
satisfaction measures the attitudes of trade customers.
Number of Complaints: The number of complaints lodged by customers in a given
time period.
2.9 Net Promoter6
Net promoter is a measure of the degree to which current customers will recommend a product, service, or company.
Net Promoter Score (I) = Percentage of Promoters (%) – Percentage of Detractors (%)
Net promoter is claimed to be a particularly useful measure of customer satisfaction and/or loyalty.
Purpose: To measure how well the brand or company is succeeding in creating satisfied, loyal customers.
Net Promoter Score7 (NPS) is a registered trademark of Frederick R. Reichheld, Bain &
Company, and Satmetrix that is a particularly simple measure of the satisfaction/loyalty
of current customers. Customers are surveyed and asked (on a ten-point scale) how
likely they are to recommend the company or brand to a friend or colleague. Based on
their answers to this single question, customers are divided into
■
52
Promoters: Customers who are willing to recommend the company to others
(who gave the company a rating of 9 or 10).
MARKETING METRICS
■
■
Passives: Satisfied but unenthusiastic customers (ratings of 7 or 8).
Detractors: Customers who are unwilling to recommend the company to
others (ratings of 0 to 6).
High NPSs generally mean that a company is doing a good job of securing their customers’ loyalty and active evangelism. Low and negative Net Promoter Scores are
important early warning signals for the firm. Because the metric is simple and easy to
understand, it provides a stable measure companies use to motivate employees and
monitor progress.
Construction
The Net Promoter Score (NPS) is created by subtracting the percentage of detractors
among current customers from the percentage of promoters among current customers.
Net Promoter Score (I) = Percentage of Promoters (%) – Percentage of Detractors (%)
For example if a survey of a company’s customers reports that there were 20% promoters, 70% passives, and 10% detractors, the company would have a Net Promoter
Score of 20–10 =10.
Data Sources, Complications and Cautions
Although the trademarked NPS asks a specific question, uses a 10-point scale, and
defines promoters, passives, and detractors in a particular way (detractors are those
giving ratings of 0 through 6), it is easy to imagine other versions of NPS that differ
with respect to the wording of the question, the scale used (1 through 5 rather than
0 through 10), and the definitions (and labels) of the resulting groups of responders.
The defining features of NPS are that it is constructed from responses to a question
about willingness to recommend and is a net measure found by subtracting the fraction
unwilling to recommend from the fraction willing to recommend and leaving out those
in the middle.
The same NPS score can indicate different business circumstances. For instance, a Net
Promoter Score of zero can indicate highly polarized customers, 50% promoters, 50%
detractors, or a totally ambivalent customer base, 100% passives. Getting the NPS score
may be a good way of starting a discussion about customer perceptions of the brand. As
it is an average of current customers’ responses, managers must drill down to the data
to understand the precise situation their business faces.
This score in specific circumstances can generate results that could mislead a manager
who is not being careful. For example, consider a company whose current customers
are 30% promoters, 30% detractors, and 40% passives. This company’s NPS is an unimpressive zero, or 30%-30%.
Chapter 2 Share of Hearts, Minds, and Markets
53
Suppose next that a new competitor steals two-thirds of the company’s detractors, and
because these detractors immediately defect to the new competitor, they cease to be
customers of the company. The NPS is remeasured.
Promoters are now 30% / (100% – 20% = 80%) = 37.5% of the customers that remain.
Passives are now 40% / (100% – 20% = 80%) = 50% of the customers that remain.
Detractors are now only (30% – 20% = 10%) / (100% – 20% = 80%) = 12.5% of the
customers that remain.
The NPS is now 37.5% – 12.5% = a very healthy looking + 25.
The defection of the most vulnerable and unhappy customers led directly to an increase
in NPS. Managers should make sure they fully understand what has happened.
While benchmarking is often a useful exercise, it is inappropriate to directly apply this
measure across categories. Some products are in categories that are more likely to gain
engagement, both positive and negative, than others.
A high Net Promoter Score, while generally desirable, does beg the question whether
the company is properly monetizing the value they are providing to the consumer. The
easiest way to develop a high Net Promoter Score is to provide a highly valued product
free to customers. Why wouldn’t they be happy to recommend you? While there may
be strategic reasons for situations like this to be acceptable to the company in the short
or medium term, this probably won’t be a viable long-term strategy.
The Net Promoter Score is calculated from survey data. As such it may suffer from the
problems common to most surveys, and the results should be interpreted in light of
other data, such as sales trends. Is increased customer satisfaction leading to increased
sales? If so, fine; if not, why not?
Although the Net Promoter Score has received much attention and relatively rapid
adoption, it has also been the target of a recent award-winning article. Consultant
Timothy Keiningham and his co-authors claim the benefits of the measure have been
overstated relative to other measures of loyalty and satisfaction.8
2.10 Willingness to Search
Although many metrics explore brand loyalty, one has been called the “acid test.”
That is,
Willingness to Search (%) = Percentage of Customers Willing to Delay Purchases,
Change Stores, or Reduce Purchase Quantities to
Avoid Switching Brands
54
MARKETING METRICS
This metric can tell a company much about the attitudes of its customers and whether
its position in the market is likely to be defensible against sustained pressure from a
competitor.
Purpose: To assess the commitment of a firm’s or a brand’s
customer base.
Brand or company loyalty is a key marketing asset. Marketers evaluate aspects of it
through a number of metrics, including repurchase rate, share of requirements, willingness to pay a price premium, and other AAU measures. Perhaps the most fundamental
test of loyalty, however, can be captured in a simple question: When faced with a situation in which a brand is not available, will its customers search further or substitute the
best available option?
When a brand enjoys loyalty at this level, its provider can generate powerful leverage
in trade negotiations. Often, such loyalty will also give providers time to respond to a
competitive threat. Customers will stay with them while they address the threat.
Loyalty is grounded in a number of factors, including
■
Satisfied and influential customers who are willing to recommend the brand.
■
Hidden values or emotional benefits, which are effectively communicated.
■
A strong image for the product, the user, or the usage experience.
Purchase-based loyalty metrics are also affected by whether a product is broadly
and conveniently available for purchase, and whether customers enjoy other options in
its category.
Construction
Willingness to Search: The likelihood that customers will settle for a second-choice
product if their first choice is not available. Also called “accept no substitutes.”
Willingness to search represents the percentage of customers who are willing to leave a
store without a product if their favorite brand is unavailable. Those willing to substitute
constitute the balance of the population.
Chapter 2 Share of Hearts, Minds, and Markets
55
Data Sources, Complications, and Cautions
Loyalty has multiple dimensions. Consumers who are loyal to a brand in the sense of
rarely switching may or may not be willing to pay a price premium for that brand or
recommend it to their friends. Behavioral loyalty may also be difficult to distinguish
from inertia or habit. When asked about loyalty, consumers often don’t know what they
will do in new circumstances. They may not have accurate recall about past behavior,
especially in regard to items with which they feel relatively low involvement.
Furthermore, different products generate different levels of loyalty. Few customers
will be as loyal to a brand of matches, for example, as to a brand of baby formula.
Consequently, marketers should exercise caution in comparing loyalty rates across
products. Rather, they should look for category-specific norms.
Degrees of loyalty also differ between demographic groups. Older consumers have been
shown to demonstrate the highest loyalty rates.
Even with these complexities, however, customer loyalty remains one of the most
important metrics to monitor. Marketers should understand the worth of their brands
in the eyes of the customer—and of the retailer.
2.11
Neuroscience Measures
(with Dr. Manuel Garcia-Garcia and Pasha Davoudian)1
In this section we discuss important methods of studying consumers made possible by
advances in neuroscience technology and outline the biometrics associated with them
that are being applied in marketing.
Electroencephalography (EEG)
Functional Magnetic Resonance Imaging (fMRI)
Facial Action Coding System (FACS)
Eye Tracking
Skin Conductance Response
Heart Rate
Electromyography (EMG)
These techniques are often used, for example, to improve the effectiveness of
advertising.
1
Thanks to Dr. Manuel Garcia-Garcia of New York University Stern School of Business and Neuroscience University
of Virginia neuroscience student Pasha Davoudian for their invaluable guidance and assistance with this section.
56
MARKETING METRICS
Purpose: Using Technology to Understand Consumer Behavior.
Technological advances have allowed marketers to gain greater insight into how consumers react, especially pre- or non-consciously. We have a better understanding of
how consumers view visual advertisements and even which parts of the consumer’s
brain are active when deciding between alternatives. Many companies provide specialist consultancy services to help marketers benefit from these new techniques.
Current research in cognitive neuroscience suggests that decision making is less
deliberate than once thought and instead relies heavily on early emotional responses.
Traditional market research methods, such as focus groups and surveys, are unable to
assess consumers’ initial, preconscious reactions. Consumer neuroscience methods can
reveal insights into early emotional responses to help us better understand these reactions and so give a unique insight into consumer behavior.
Construction
We consider four of the most important measures used in trying to understand consumer reactions. These are classified here by the technology rather than the measure, as
the choice of technology is generally the way that marketers will encounter the metrics.
Electroencephalography (EEG)
An EEG is a technique that involves taking surface readings from electrodes embedded
in headgear. The electrodes pick up voltage fluctuations within the brain and can pin
down activity to an area of the brain.
EEG Units of Measurement: An EEG measures brain waves in terms of Hertz
(frequency) and Micro-voltage (amplitude).
EEGs are useful for exploring such things as, “what is the consumer’s immediate
response to an advertisement?” EEGs allow a better understanding through rapid monitoring of general brain activity. This removes the subjectivity from many assessments
of marketing: i.e., we don’t need to rely on surveys of consumer attitudes.
This approach can assess the consumer’s attentional effort. This captures two different ideas. 1) Top-Down attention, voluntary attentional focus, and 2) Bottom-Up
attention, involuntary responses often to the novel, rewarding, or threatening. When
properly interpreted, marketing applications include giving insight into branding and
advertising effects. In advertising an EEG can provide a measure of attention from
the changes in certain brain-wave patterns over regions of the brain. The strength of
brands, and the type of associations held about the brand, can also be assessed by the
activity the brand generates in the consumer’s brain.
Chapter 2 Share of Hearts, Minds, and Markets
57
Functional Magnetic Resonance Imaging (fMRI)
fMRI involves the consumer entering a scanning machine, a large machine that is
non-portable and expensive. Although the technique is non-invasive, it is an artificial
environment, and entering the scanner presents challenges to those who suffer from
claustrophobia.
The consumer will usually be shown stimuli and their brain responses monitored. This
process can be effective at visualizing brain processes at fine spatial resolution.
fMRI Measurement: Blood oxygen level dependent (BOLD) brain tracking that
resolves activity to the micrometer scale per second.
When showing the results, the intensity of activity in the consumer’s brain is color
coded to indicate relative brain activity. Red normally depicts high brain activity and
blue represents lower brain activity. Thus, we see which areas of the brain were most
active when seeing a certain stimuli. A researcher might then talk about areas associated with emotional reactions being more active when the consumer was shown one
brand rather than another. One interesting example is the “Pepsi Challenge” study. This
identified brain regions that attracted greater blood flow when brand associations were
triggered. (See the McClure et al. and Knutson et al. references in the further reading on
this case and another key study on the role of the reward system in predicting consumer
behavior.)
Facial Action Coding System (FACS)
Facial coding involves attempting to identify the consumer’s mood/reaction through
his or her facial movements. There is no single unit of measurement. The system
employs numerous facial action unit characterizations along with related general head
and specific sub facial movements. Users are trained to identify these. Automatic classification of observable facial expressions can be done by several software programs,
some of which can be implemented through webcams, with more or less reliable results.
This technique may help to identify a person’s underlying mood or their reaction to a
stimulus—reactions that otherwise might not be explicitly expressed, such as the subtle
facial movements that correlate to particular mindsets or attitudes. Facial coding can be
used to analyze consumer reactions to proposed product features or to choose between
different versions of an advertisement. One can consider how consumers respond emotionally to a brand or even use the technique as part of an autopsy on why a marketing
campaign failed.
An example of facial coding being used in marketing is the work of Thales Teixeira
and his colleagues. These researchers examined the surprise and joy of consumers who
were viewing a series of online advertisements. They used facial coding software to
fit a virtual facial mask on the consumers who were watching the advertisements and
58
MARKETING METRICS
measured deviations from the baseline to measure the emotions being experienced. For
example, smiles were detected from deviations related to the corners of the lips. The
emotions were then linked to what the consumer was paying attention to at the time.
The emotions were also linked to whether the consumers clicked past a given advertisement or watched it. Using such an approach a marketer can get a clearer view of the
emotions consumers experience when seeing an advertisement and so design advertisements to better gain the consumers’ attention and sustain the consumers’ interest.
Eye Tracking
Eye tracking provides a real-time record of where visual attention is directed. This can
also show how pupil dilation changes. This measure is a useful indicator of emotional
arousal.
Small, high-resolution video cameras are placed near, but without obstructing, the
consumer’s eyes. A small non-invasive light guides the camera to track what is being
viewed while other cameras capture related information. Typically studies are run on
stationary subjects but additional hardware can be employed to track a subject’s gaze
when moving.
Eye Tracking Measures: Fixations per second, saccades (eye movements), pupil
size, blinks per second.
Fixations per second when a person is viewing a stimuli can help to indicate how much
attention the consumer is paying to, for example, a sample advertisement. In product
and package design eye tracking can test what the consumer is paying visual attention
to and gives an idea of the consumer’s emotional arousal from pupil dilation. In shopper marketing eye tracking can map out where the consumers look in a shopping aisle,
on a shelf, or when examining individual products. Combining eye tracking with shelf
plans (planograms) can yield useful insights into how consumers visually search a shelf
to find and select products. Such studies can provide a wealth of practical information
about the impact of different shelf configurations on product search and selection.
Applying These Measures to Marketing Problems
We borrow an example from Trabulsi, Garcia-Garcia, and Smith. They describe a situation where a marketer tested an advertisement using traditional marketing research
methods but found that it was unlikely to be effective. The basic approach still seemed
promising and so a decision was taken to recreate the advertisement but in a more
effective manner. The marketer, therefore, wanted to know exactly, i.e., to the second,
what was failing. Measuring consumer reactions to an advertisement with an EEG the
marketers could precisely find out which seconds of the advertisement were strong and
which weak in terms of consumer attention, memory, and emotional reaction. The
Chapter 2 Share of Hearts, Minds, and Markets
59
marketer could also see which elements of the advertisement were especially interesting
and/or confusing by tracking at which seconds when watching the advertisement high
demands were being placed on the consumers’ cognitive capacities.
Other Important Measures Related To Neuro-Marketing
This guide is not designed to be a specialist neuro-marketing publication so we have
focused on a few important technologies. Within this field, however, there are other
measures taken, often simultaneously, with those above. We might want to know how
the consumer’s body is reacting; are they excited? Preparing for action? Just bored?
Electromyography (EMG) measures facial micro-muscle movements, below the level of
observable expressions, involuntarily associated with emotional reactions, such as the
“frown” and “smile” muscles.
Electrodermal activity (EDA) is electrical current passing through the skin. It is a function of the amount of perspiration on the skin. Perspiration conducts electricity and is
a sign of emotional arousal which is measured by Galvanic Skin Response (GSR) and
Skin Conductance Response (SCR).
Heart rate can slow down when attention increases and speeds up when experiencing
emotional arousal. Respiration measures record how deep and fast a person breathes
which is associated with arousal.
Response latencies (time delays) measure how fast we respond to choice and judgment
tasks. Quicker action is usually assumed to show a stronger mental association between
concepts in long-term memory. Two words or images are presented in rapid succession. When the second item appears, a choice or judgment has to be made as rapidly as
possible. Three main types of behavioral response testing are semantic priming, affective priming, and the Implicit Association Test (IAT).
Why Use These Techniques?
Brand-equity measures often capture the progress of a brand’s marketplace performance, but neuro-marketing measures have potential to tell us more about the consumers’ experiences. We can also hope to better understand non-conscious consumer
reactions, which consumers cannot tell us about in surveys even if they wish to do so.
Where these are likely to be especially useful is where the upfront costs are modest
compared to the ongoing marketing costs. Here even a modest increase in understanding can mean large benefits from the more efficiently deployed ongoing costs. A case in
point may be advertising where typically media buys, time on TV, is a large proportion
of the costs. (The cost to create a commercial is often modest compared to the cost to
air it on network TV.) Anything that even marginally improves the effectiveness of an
60
MARKETING METRICS
advertisement can be highly worthwhile. (For more on discussion of advertising testing,
see Section 12.3 on the Gross model.)
EXAMPLE: Heritage Inc. has created a 30-second advertisement that is quite effective; for
every dollar spent there was an expected contribution of $1.50. (This comes from the company’s
internal assessment of past advertising.)
After a study a neuro-marketing consultancy, Electrode Partners, recommended heavy edits to
the advertisement. A researcher used a combination of eye-tracking and brain activity measures to
assess which sections effectively engaged viewers. By eliminating the sections that were less effective
and/or confusing, the ad was shortened. This sacrificed a small amount of effectiveness; each spot
would be only 80% as effective as the 30-second advertisement, but the new spots would last only 15
seconds. Media costs for a 30-second spot are twice that for a 15-second spot so although each spot
would only have 80% of the impact each spot would be half the cost.
Heritage is committed to their media spend given their agreement with the TV network. Heritage
will, therefore, buy twice as many 15-second spots as they would 30-second spots and the company
treats media spend as a fixed cost. Electrode’s study plus making the necessary edits to the advertisement cost a total of $250,000. In Table 2.9 we can see that the return on investment is quite substantial for the study. (For more on Return on Investment see Section 11.2.)
Table 2.9 Return On Investment For Study
All Dollar Terms in $000s
Baseline
After Study
Effectiveness of 30 Sec./15 Sec. Spot (A)
Given
100%
80%
Relative # of Spots 30 Sec./15 Sec. (B)
Given
1
2
Relative Effectiveness (C)
A*B
100%
160%
Expected Lift From 30 Sec. Spot (D)
Given
1.5
1.5
Expected Lift From Media Spend (E)
C*D
1.5
2.4
Media Spend (F)
Given
$1,000
$1,000
Contribution From Media Spend (G)
E*F
$1,500
$2,400
Profit From Media Spend (H)
G-F
$500
$1,400
Study and Advertisement Change Costs (I)
Given
$250
Chapter 2 Share of Hearts, Minds, and Markets
61
Profit Without Study (J)
H
$500
Profit With Study (K)
H-I
$1,150
Profit Change From Study (L)
K-J
$650
ROI of Study (M)
L/I
260%
The study’s success is driven by the fact that the study’s upfront cost is less than
the ongoing benefits gained by more efficient deployment of the media spend. The
15-second advertisements give much more bang for the buck than the 30-second
advertisements.
The benefits from small improvements in effectiveness get even greater when the ongoing spending increases relative to the initial costs. See Table 2.10 for the benefits that
would have been gained had Heritage been committed to five times more media spend.
Table 2.10 Return On Investment For Study When Media Spend Is Higher
All Dollar Terms in $000s
Baseline
After Study
Media Spend is 5x Higher (F')
Given
$5,000
$5,000
Contribution From Spots (G')
E*F'
$7,500
$12,000
Profit Without Study (J')
G'-F'
$2,500
Profit With Study (K')
G'-F'-I
$6,750
Profit Change From Study (L')
K'-J'
$4,250
ROI of Study (M')
L'/I
1700%
Data Sources, Complications and Cautions
One of the main limitations of these techniques is that the machinery can be cumbersome and so hard or impossible to deploy in everyday situations. While we can ask consumers to participate in these tests their reactions will, to a certain extent, be influenced
by the setting they are in. Participating in an experiment in an fMRI scanner is not (yet)
the same as shopping in your local grocer even if some of the measuring devices are
rapidly becoming smaller and more comfortable. As technology for gathering neuroscience metrics becomes less obtrusive we might expect some of these problems to be
alleviated.
62
MARKETING METRICS
Compared to other marketing metrics, neuroscience requires great expertise and training in the underlying science to interpret the data. Further, although companies are
accumulating studies that assess reliability and validity of the technologies and metrics,
much remains to be done.
As Varan and his colleagues note the need to assess the variance and reliability of
neuro-metrics within and between firms has become increasingly apparent. There
remains considerable variation among suppliers and uncertainty about what differences are due to lack of reliability versus what differences are driven by each firm’s
unique collection methods.
Our understanding of the brain has increased considerably in recent years but it still
remains limited. Scientists will talk of areas of the brain that are associated with a certain feeling or activity. This does not mean that there is a specific spot in the brain that
if we find it we can tell exactly your brand preference. Neither can we hope to understand a consumer’s precise willingness to pay for any good. These techniques may be
useful in understanding how consumers are likely to behave, and so may help at an
early stage of the marketing process, but we would still recommend using traditional
metrics to monitor the performance of your marketing campaigns.
Further Reading
Mahzarin R. Banaji and Anthony G. Greenwald, (2013), Blindspot: Hidden Biases of Good
People, Delacorte Press.
Kirk Hendrickson, Kusum L. Ailawadi (2014), Six Lessons for In-Store Marketing from Six
Years of Mobile Eye-Tracking Research, in Dhruv Grewal, Anne L. Roggeveen, Jens NordfÄlt
(ed.) Shopper Marketing and the Role of In-Store Marketing (Review of Marketing Research,
Volume 11) Emerald Group Publishing Limited, pp. 57–74.
Rick B. Knutson, S., G. E. Wimmer, D. Prelec, & G. Loewenstein (2007). Neural predictors of
purchases. Neuron, 53(1), 147-156.
Samuel M. McClure, Jian Li1, Damon Tomlin, Kim S. Cypert, Latané M. Montague, P. Read
Montague (2004) Neural Correlates of Behavioral Preference for Culturally Familiar Drinks,
Neuron, 44 (2), 379-387.
Thales Teixeira, Michel Wedel, and Rik Pieters, (2012), Emotion-Induced Engagement in
Internet Video Advertisements, Journal of Marketing Research, Vol. XLIX (April), 144 –159.
Julia Trabulsi, Manuel Garcia-Garcia, Michael E. Smith, (Draft Paper) Consumer Neuroscience:
A Method for Optimizing Marketing Communication
Duane Varan, Annie Lang, Patrick Barwise, Rene Weber, and Steven, Bellman. (2015). How
Reliable Are Neuromarketers’ Measures of Advertising Effectiveness: Data from Ongoing
Chapter 2 Share of Hearts, Minds, and Markets
63
Research Holds No Common Truth among Vendors. Journal of Advertising Research, 55(2),
176–191.
Vinod Venkatraman, Angelika Dimoka, Paul A. Pavlou, Khoi Vo, William Hampton, Bryan
Bollinger, Hal E. Hershfield, Masakazu Ishihara, and Russell S. Winer, (In Press) Predicting
Advertising Success Beyond Traditional Measures: New Insights from Neurophysiological
Methods and Market Response Modeling, Journal of Marketing Research.
Joe Willke and Blake Burrus (2013) Making Advertising More Of A Science Than An Art,
What’s Next, Volume 1 No. 2, The Nielsen Company.
64
MARKETING METRICS
Table 2.11
Summary Of Neuro-Metrics
Measurement Unit
Technological Specifications
Targeted Processes
Potential Marketing Uses
Hertz (frequency)
and Micro-voltage
(amplitude) together
define characteristic
wave forms.
Surface electrodes arranged
in headgear (cap, helmet, etc.)
are painless and relatively
comfortable. Costs associated
with EEGs can range from
$500 (single scan) to more
than $10,000 for the set-up of
an entire EEG hardware and
software set.
Well suited to explore general
cognitive decision making,
initial responses, problem
solving, etc. due to the high
temporal resolution of the
datasets acquired.
Branding: Identifies brain-wave
patterns that emerge when
presented words or images
that are strongly associated in
memory.
65
Blood oxygen level
dependent (BOLD)
brain tracking
resolves activity to the
micrometer scale per
second.
Voxel (3D-pixel and
volume) intensity
is color coded. Red
normally depicts high
brain activity, blue
lower brain activity.
Two magnetic fields stabilize
and excite brain nuclei and
measure the resulting changes
via magnetic coils. fMRIs range
from $500-800 for a single scan
to $1-3 million for entire setups.
Can pinpoint brain regions
corresponding with behavior.
Best for attempting to find
brain regions responsible for
sustained behaviors such as food
or drink seeking, judgment, etc.
Branding: Identify brain regions
that attract greater blood flow
when strong associations are
triggered.
User classifies
movements, features
into action units
(AUs). Relative
intensities used to
indirectly assess mood
or intentions.
No standard unit.
Employs more
than 50 facial AU
characterizations
along with another
50 general head and
specific sub facial
movements.
Users are trained to accurately
identify the various facial
characterizations (50+).
Automating AU classification
can be done via facial recognition
software.
Well suited to identify a
person’s underlying mood or
intentions, which may otherwise
not be explicitly expressed.
Product Testing: Analyze
customer reactions to proposed
product features.
Fixations per second,
saccade linear
mapping, blinks per
second, voxels (for
attentiveness heat
maps).
Small, high-resolution video
cameras placed near the eyes
use small non-invasive light to
track the view. In parallel other
cameras capture the scene.
Hardware and software packages
vary greatly and can cost
anywhere from hundreds to tens
of thousands of dollars.
Facial Action Coding
System (FACS)
Voltage fluctuations of
thousands of neurons
per millisecond at a
spatial resolution of up
to a single centimeter.
Eye Tracking
Chapter 2 Share of Hearts, Minds, and Markets
Electroencephalography
(EEG)
Metric/Measure
Functional
Magnetic
Resonance
Imaging (fMRI)
Technology
Video recordings
track pupil via
(near) infrared light.
Additional hardware
can track gaze when
moving.
Advertising: Measure of
attention, changes in brain-wave
patterns.
Advertising testing: Choose
between ad versions or
determine why ad campaigns
failed.
Brand emotions: measures
responses to brand.
Useful for investigating the
visual system, particularly the
length and order in which a
subject views various aspects of
a scene.
Product, Package Design: test
attention/arousal.
Advertising: Measure the
number of fixations per second
(fps) viewing an ad.
Shopper marketing: measures
where consumers look in aisle,
on shelf, or on product.
This page intentionally left blank
3
MARGINS AND PROFITS
Introduction
Key concepts covered in this chapter:
Margins
Selling Prices and Channel Margins
Average Price per Unit and Price
per Statistical Unit
Variable Costs and Fixed Costs
Marketing Spending—Total, Fixed, and
Variable
Break-Even Analysis and Contribution
Analysis
Target Volume
Peter Drucker has written that the purpose of a business is to create a customer. As
marketers, we agree. But we also recognize that a business can’t survive unless it makes
a margin as well as a customer. At one level, margins are simply the difference between
a product’s price and its cost. This calculation becomes more complicated, however,
when multiple variations of a product are sold at multiple prices, through multiple
channels, incurring different costs along the way. For example, a recent Business Week
article noted that less “than two-thirds of GM’s sales are retail. The rest go to rental-car
agencies or to company employees and their families—sales that provide lower gross
margins.”1 Although it is still the case that a business can’t survive unless it earns a positive margin, it can be a challenge to determine precisely what margin the firm actually
does earn.
In the first section of this chapter, we’ll explain the basic computation of unit and
percentage margins, and we’ll introduce the practice of calculating margins as a
percentage of selling price.
Next, we’ll show how to “chain” this calculation through two or more levels in a
distribution channel and how to calculate end-user purchase price on the basis of a
67
marketer’s selling price. We’ll explain how to combine sales through different channels
to calculate average margins and how to compare the economics of different distribution
channels.
In the third section, we’ll discuss the use of “statistical” and standard units in tracking
price changes over time.
We’ll then turn our attention to measuring product costs, with particular emphasis
on the distinction between fixed and variable costs. The margin between a product’s
unit price and its variable cost per unit represents a key calculation. It tells us how
much the sale of each unit of that product will contribute to covering a firm’s fixed
costs. “Contribution margin” on sales is one of the most useful marketing concepts. It
requires, however, that we separate fixed from variable costs, and that is often a challenge. Frequently, marketers must take “as a given” which of their firm’s operating
and production costs are fixed and which are variable. They are likely, however, to be
responsible for making these fixed versus variable distinctions for marketing costs. That
is the subject of the fifth section of this chapter.
In the sixth section, we’ll discuss the use of fixed- and variable-cost estimates in calculating the break-even levels of sales and contribution. Finally, we’ll extend our calculation of break-even points, showing how to identify sales and profit targets that are
mutually consistent.
Metric
Construction
Considerations
Purpose
3.1
Unit Margin
Unit price less the
unit cost.
What are the standard
units in the industry?
May not reflect
contribution margin
if some fixed costs are
allocated.
Determine value of
incremental sales.
Guide pricing and
promotion.
3.1
Margin (%)
Unit margin as a
percentage of unit
price.
May not reflect
contribution margin
if some fixed costs are
allocated.
Compare margins
across different
products/sizes/
forms of product.
Determine value of
incremental sales.
Guide pricing
and promotion
decisions.
68
MARKETING METRICS
Metric
Construction
Considerations
Purpose
3.2
Channel
Margins
Channel profits
Distinguish margin
as percentage of
on sales (usual) from
channel selling price. markup on cost (also
encountered).
Evaluate channel
value added in
context of selling
price. Calculate
effect of price
changes at one
level of channel on
prices and margins
at other levels in
the same channel
(supply chain).
3.3
Average Price
per Unit
Can be calculated as Some units may have
total revenue divided greater relevance from
by total unit sales.
producers’ perspective
than consumers’ (e.g.,
ounces of shampoo vs.
bottles). Changes may
not be result of pricing
decisions.
Understand how
average prices are
affected by shifts in
pricing and product
mix.
3.3
Price per
Statistical Unit
SKU prices
weighted by relevant
percentage of each
SKU in a statistical
unit.
Percentage SKU mix
should correspond
over medium-term to
actual mix of sales.
Isolate effect of
price changes
from mix changes
by standardizing
the SKU mix of a
standard unit.
3.4
Variable and
Fixed Costs
Divide costs into two
categories: those that
vary with volume
(variable) and those
that do not (fixed).
Variable costs may
include production,
marketing, and selling
expenses. Some
variable costs depend
on units sold; others
depend on revenue.
Understand how
costs are affected
by changes in sales
volume.
3.5
Marketing
Spending
Analyze costs that
comprise marketing
spending.
Can be divided into
fixed and variable
marketing costs.
Understand how
marketing spending
changes with sales.
Continues
Chapter 3 Margins and Profits
69
Metric
Construction
Considerations
Purpose
3.6
Contribution per
Unit
Unit price less
unit variable cost.
Ensure that
marketing
variable costs have
not already been
deducted from
price.
3.6
Contribution
Margin (%)
Contribution per
unit divided by
unit price.
Ensure that
variable costs
are consistently
based on units
or revenue, as
appropriate.
3.6
Break-Even Sales
Level
For unit breakeven, divide
fixed costs by
contribution per
unit. For revenue
break-even,
divide fixed costs
by contribution
margin (%).
Variable and fixed
cost estimates
may be valid
only over certain
ranges of sales and
production.
Rough
indicator of
project
attractiveness and
ability to earn
profit.
3.7
Target Volume
Adjust breakeven calculation
to include profit
target.
Variable
marketing costs
must be reflected
in contribution
margins. Sales
increases often
require increased
investment or
working capital.
Ensure that unit
sales objectives
will enable firm to
achieve financial
hurdle rates for
profit, ROS, or
ROI.
3.7
Target Revenues
Convert target
volume to target
revenues by using
average prices per
unit. Alternatively,
combine cost
and target data
with knowledge
of contribution
margins.
70
MARKETING METRICS
Same as above.
Understand profit
impact of changes
in volume.
Calculate breakeven level of sales.
Same as above, but
applies to dollar
sales.
Same as above,
applied to revenue
objectives.
3.1 Margins
Margin (on sales) is the difference between selling price and cost. This difference is
typically expressed either as a percentage of selling price or on a per-unit basis.
Unit Margin ($) = Selling Price per Unit ($) – Cost per Unit ($)
Margin (%) =
Unit Margin ($)
Selling Price per Unit ($)
Managers need to know margins for almost all marketing decisions. Margins represent a key factor in pricing, return on marketing spending, earnings forecasts, and
analyses of customer profitability.
Purpose: To determine the value of incremental sales,
and to guide pricing and promotion decisions.
Margin on sales represents a key factor behind many of the most fundamental business
considerations, including budgets and forecasts. All managers should, and generally
do, know their approximate business margins. Managers differ widely, however, in the
assumptions they use in calculating margins and in the ways they analyze and communicate these important figures.
Percentage Margins and Unit Margins: A fundamental variation in the way people
talk about margins lies in the difference between percentage margins and unit margins
on sales. The difference is easy to reconcile, and managers should be able to switch back
and forth between the two.
What is a unit? Every business has its own notion of a “unit,” ranging from a ton of
margarine, to 64 ounces of cola, to a bucket of plaster. Many industries work with
multiple units and calculate margin accordingly. The cigarette industry, for example,
sells “sticks,” “packs,” “cartons,” and 12M “cases” (which hold 1,200 individual cigarettes). Banks calculate margin on the basis of accounts, customers, loans, transactions,
households, and branch offices. Marketers must be prepared to shift between such
varying perspectives with little effort because decisions can be grounded in any of these
perspectives.
Construction
Unit Margin ($) = Selling Price per Unit ($) – Cost per Unit ($)
Margin (%) =
Unit Margin ($)
Selling Price per Unit ($)
Chapter 3 Margins and Profits
71
Percentage margins can also be calculated using total sales revenue and total costs.
[Total Sales Revenue ($) – Total Cost ($)]
Margin (%) =
Total Sales Revenue ($)
When working with either percentage or unit margins, marketers can perform a simple
check by verifying that the individual parts sum to the total.
To Verify a Unit Margin ($): Selling Price per Unit = Unit Margin + Cost per Unit
To Verify a Margin (%): Cost as % of Sales = 100% – Margin %
EXAMPLE: A company markets sailcloth by the lineal yard. Its cost basis and selling price for
standard cloth are as follows:
Unit Selling Price (Selling Price per Unit) = $24 per Lineal Yard
Unit Cost (Cost per Unit) = $18 per Lineal Yard
To calculate unit margin, we subtract the cost from the selling price:
Unit Margin = $24 per Yard – $18 per Yard
= $6 per Yard
To calculate the percentage margin, we divide the unit margin by the selling price:
Margin (%) =
=
($24 – $18) per Yard
$24
$6
$24
= 25%
Let’s verify that our calculations are correct:
Unit Selling Price = Unit Margin + Unit Cost
$24 per Yard = $6 per Yard + $18 per Yard
correct
A similar check can be made on our calculations of percentage margin:
100% – Margin on Sales (%) = Cost as % of Selling Price
100% – 25% =
$18
$24
75% = 75%
72
MARKETING METRICS
correct
When considering multiple products with different revenues and costs, we can calculate overall margin (%) on either of two bases:
■
■
Total revenue and total costs for all products, or
The dollar-weighted average of the percentage margins of the different
products
EXAMPLE: The sailcloth company produces a new line of deluxe cloth, which sells for $64 per
lineal yard and costs $32 per yard to produce. The margin on this item is 50%.
Unit Margin ($) = $64 per Yard – $32 per Yard
= $32 per Yard
Margin (%) =
=
($64 – $32)
$64
$32
$64
= 50%
Because the company now sells two different products, its average margin can only be calculated
when we know the volume of each type of goods sold. It would not be accurate to take a simple average of the 25% margin on standard cloth and the 50% margin on deluxe cloth, unless the company
sells the same dollar volume of both products.
If, one day, the company sells 20 yards of standard cloth and two yards of deluxe cloth, we can calculate its margins for that day as follows (see also Table 3.1):
Total Sales = 20 Yards at $24, and 2 Yards at $64
= $608
Total Costs = 20 Yards at $18, and 2 Yards at $32
= $424
Margin ($) = $184
Margin (%) =
Margin ($184)
Total Sales ($608)
= 30%
Because dollar sales differ between the two products, the company margin of 30% is not a simple
average of the margins of those products.
Chapter 3 Margins and Profits
73
Table 3.1 Sales, Costs, and Margins
Standard
Deluxe
Total
Sales in Yards
20
2
22
Selling Price per Yard
Total Sales $
$24.00
$480.00
$64.00
$128.00
$608.00
Cost per Yard
Total Costs $
$18.00
$360.00
$32.00
$64.00
$424.00
Total Dollar Margin ($)
Unit Margin
$120.00
$6.00
$64.00
$32.00
$184.00
$8.36
Margin (%)
25%
50%
30%
Data Sources, Complications, and Cautions
After you determine which units to use, you need two inputs to determine margins:
unit costs and unit selling prices.
Selling prices can be defined before or after various “charges” are taken: Rebates,
customer discounts, brokers’ fees, and commissions can be reported to management
either as costs or as deductions from the selling price. Furthermore, external reporting can vary from management reporting because accounting standards might dictate
a treatment that differs from internal practices. Reported margins can vary widely,
depending on the calculation technique used. This can result in deep organizational
confusion on as fundamental a question as what the price of a product actually is.
Please see Section 8.4 on price waterfalls for cautions on deducting certain discounts
and allowances in calculating “net prices.” Often, there is considerable latitude on
whether certain items are subtracted from list price to calculate a net price or are added
to costs. One example is the retail practice of providing gift certificates to customers
who purchase certain amounts of goods. It is not easy to account for these in a way
that avoids confusion among prices, marketing costs, and margins. In this context, two
points are relevant: (1) Certain items can be treated either as deductions from prices or
as increments to cost, but not both. (2) The treatment of such an item will not affect the
unit margin, but will affect the percentage margin.
Margin as a percentage of costs: Some industries, particularly retail, calculate margin as a percentage of costs, not of selling prices. Using this technique in the previous
example, the percentage margin on a yard of standard sailcloth would be reckoned as
74
MARKETING METRICS
the $6.00 unit margin divided by the $18.00 unit cost, or 33%. This can lead to confusion. Marketers must become familiar with the practices in their industry and stand
ready to shift between them as needed.
Markup or margin? Although some people use the terms “margin” and “markup”
interchangeably, this is not appropriate. The term “markup” commonly refers to the
practice of adding a percentage to costs in order to calculate selling prices.
To get a better idea of the relationship between margin and markup, let’s calculate a
few. For example, a 50% markup on a variable cost of $10 would be $5, yielding a retail
price of $15. By contrast, the margin on an item that sells at a retail price of $15 and that
carries a variable cost of $10 would be $5/ $15, or 33.3%. Table 3.2 shows some common margin/markup relationships.
Table 3.2 Relationship Between Margins and Markups
Price
Cost
Margin
Markup
$10
$9.00
10%
11%
$10
$7.50
25%
33%
$10
$6.67
33.3%
50%
$10
$5.00
50%
100%
$10
$4.00
60%
150%
$10
$3.33
66.7%
200%
$10
$2.50
75%
300%
One of the peculiarities that can occur in retail is that prices are “marked up” as a percentage of a store’s purchase price (its variable cost for an item) but “marked down”
during sales events as a percentage of retail price. Most customers understand that a
50% “sale” means that retail prices have been marked down by 50%.
EXAMPLE: An apparel retailer buys t-shirts for $10 and sells them at a 50% markup. As noted
previously, a 50% markup on a variable cost of $10 yields a retail price of $15. Unfortunately, the
goods don’t sell, and the store owner wants to sell them at cost to clear shelf space. He carelessly asks
a sales assistant to mark the goods down by 50%. This 50% markdown, however, reduces the retail
price to $7.50. Thus, a 50% markup followed by a 50% markdown results in a loss of $2.50 on each
unit sold.
Chapter 3 Margins and Profits
75
It is easy to see how confusion can occur. We generally prefer to use the term margin to
refer to margin on sales. We recommend, however, that all managers clarify with their
colleagues what is meant by this important term.
EXAMPLE: A wireless provider sells a handset for $100. The handset costs $50 to manufacture
and includes a $20 mail-in rebate. The provider’s internal reports add this rebate to the cost of
goods sold. Its margin calculations therefore run as follows:
Unit Margin ($) = Selling Price – Cost of Goods Sold and Rebate
= $100 – ($50 + $20) = $30
Margin (%) =
$30
$100
= 30%
Accounting standards mandate, however, that external reports deduct rebates from sales revenue
(see Table 3.3). Under this construction, the company’s margin calculations run differently and
yield a different percentage margin:
Unit Margin ($) = Selling Price, Net of Rebate – Cost of Goods Sold
= ($100 – $20) – $50 = $30
Margin (%) =
$30
($100 – $20)
$30
=
Table 3.3
$80
= 37.5%
Internal and External Reporting May Vary
Internal Reporting
External Reporting
Dollars Received from Customer
Rebate
Sales
$100
—
$100
$100
$20
$80
Manufacturing Cost
Rebate
Cost of Goods Sold
$50
$20
$70
$50
—
$50
Unit Margin ($)
Margin (%)
$30
30.0%
$30
37.5%
In this example, managers add the rebate to cost of goods sold for the sake of internal reports. In
contrast, accounting regulations require that the rebate be deducted from sales for the purpose of
76
MARKETING METRICS
external reports. This means that the percentage margin varies between the internal and external
reports. This can cause considerable angst within the company when quoting a percentage margin.
As a general principle, we recommend that internal margins follow formats mandated
for external reporting in order to limit confusion.
Various costs may or may not be included: The inclusion or exclusion of costs generally
depends on the intended purpose of the relevant margin calculations. We’ll return to this
issue several times. At one extreme, if all costs are included, then margin and net profit
will be equivalent. On the other hand, a marketer may choose to work with “contribution
margin” (deducting only variable costs), “operating margin,” or “margin before marketing.” By using certain metrics, marketers can distinguish fixed from variable costs and
can isolate particular costs of an operation or of a department from the overall business.
Related Metrics and Concepts
Gross Margin: This is the difference between revenue and cost before accounting for
certain other costs. Generally, it is calculated as the selling price of an item, less the
cost of goods sold (production or acquisition costs, essentially). Gross margin can be
expressed as a percentage or in total dollar terms. If the latter, it can be reported on a
per-unit basis or on a per-period basis for a company.
3.2 Prices and Channel Margins
Channel margins can be expressed on a per-unit basis or as a percentage of selling
price. In “chaining” the margins of sequential distribution channels, the selling price
of one channel member becomes the “cost” of the channel member for which it
serves as a supplier.
Supplier Selling Price ($) = Customer Selling Price ($) – Customer Margin ($)
Customer Selling Price ($) =
Supplier Selling Price ($)
[1 – Customer Margin (%)]
When there are several levels in a distribution chain—including a manufacturer, distributor, and retailer, for example—one must not simply add all channel margins as
reported in order to calculate “total” channel margin. Instead, use the selling prices
at the beginning and end of the distribution chain (that is, at the levels of the manufacturer and the retailer) to calculate total channel margin. Marketers should be able
to work forward from their own selling price to the consumer’s purchase price and
should understand channel margins at each step.
Chapter 3 Margins and Profits
77
Purpose: To calculate selling prices at each level in the
distribution channel.
Marketing often involves selling through a series of “value-added” resellers. Sometimes,
a product changes form through this progression. At other times, its price is simply
“marked up” along its journey through the distribution channel (see Figure 3.1).
In some industries, such as imported beer, there may be as many as four or five channel
members that sequentially apply their own margins before a product reaches the consumer. In such cases, it is particularly important to understand channel margins and
pricing practices in order to evaluate the effects of price changes.
Buys Raw
Materials for
$0.50
Manufacturer
Buys from
Manufacturer
for $1.00
Sells to
Distributor
for $1.00
Distributor
Buys from
Distributor
for $2.00
Sells to
Wholesaler
for $2.00
Wholesaler
Sells to
Retailer
for $3.00
Unit
Margin
Margin
%
Buys from
Wholesaler
for $3.00
Retailer
Buys
from
Retailer
for $5.00
Sells to
Consumer
for $5.00
Consumer
$5.00
$0.50
$1.00
$1.00
$2.00
50%
50%
33.3%
40%
Margin ($) for entire chain $4.50
Margin (%)
90%
Figure 3.1 Example of a Distribution Channel
Remember:
78
MARKETING METRICS
Selling Price = Cost + Margin
Construction
First, decide whether you want to work “backward,” from customer selling prices to
supplier selling prices, or “forward.” We provide two equations to use in working backward, one for dollar margins and the other for percentage margins:
Supplier Selling Price ($) = Customer Selling Price ($) – Customer Margin ($)
Supplier Selling Price ($) = Customer Selling Price ($) * [1 – Customer Margin (%)]
EXAMPLE: Aaron owns a small furniture store. He buys BookCo brand bookcases from a local
distributor for $200 per unit. Aaron is considering buying directly from BookCo, and he wants
to calculate what he would pay if he received the same price that BookCo charges his distributor.
Aaron knows that the distributor’s percentage margin is 30%.
The manufacturer supplies the distributor. That is, in this link of the chain, the manufacturer is the
supplier, and the distributor is the customer. Thus, because we know the customer’s percentage
margin, in order to calculate the manufacturer’s price to Aaron’s distributor, we can use the second
of the two previous equations.
Supplier Selling Price ($) = Customer Selling Price ($) * [1 – Customer Margin (%)]
= $200 * 70% = $140
Aaron’s distributor buys each bookcase for $140 and sells it for $200, earning a margin of $60 (30%).
Although the previous example may be the most intuitive version of this formula, by
rearranging the equation, we can also work forward in the chain, from supplier prices
to customer selling prices. In a forward-looking construction, we can solve for the
customer selling price, that is, the price charged to the next level of the chain, moving
toward the end consumer.2
Customer Selling Price ($) =
Supplier Selling Price ($)
[1 – Customer Margin (%)]
Customer Selling Price ($) = Supplier Selling Price ($) + Customer Margin ($)
EXAMPLE: Clyde’s Concrete sells 100 cubic yards of concrete for $300 to a road construction
contractor. The contractor wants to include this in her bill of materials, to be charged to a local government (see Figure 3.2). Further, she wants to earn a 25% margin. What is the contractor’s selling
price for the concrete?
Chapter 3 Margins and Profits
79
Supplier to
Supplier to
Clyde
The Contractor
Customer of
Local Government
Customer of
Figure 3.2 Customer Relationships
This question focuses on the link between Clyde’s Concrete (supplier) and the contractor (customer). We know the supplier’s selling price is $300 and the customer’s intended margin is 25%.
With this information, we can use the first of the two previous equations.
Supplier Selling Price
Customer Selling Price =
(1 – Customer Margin %)
$300
=
=
(1 – 25%)
$300
75%
= $400
To verify our calculations, we can determine the contractor’s percentage margin, based on a selling
price of $400 and a cost of $300.
(Customer Selling Price – Supplier Selling Price)
Customer Margin =
Customer Selling Price
($400 – $300)
=
=
$400
$100
$400
= 25%
First Channel Member’s Selling Price: Equipped with these equations and with
knowledge of all the margins in a chain of distribution, we can work all the way back to
the selling price of the first channel member in the chain.
First Channel Member’s Selling Price ($) = Last Channel Member’s Selling Price ($) * [1 – Last
Channel Margin (%)] * [1 – Next-to-last Channel
Margin (%)] * [1 – Next-to-next-to-last Channel
Margin (%)] . . . and so on
80
MARKETING METRICS
EXAMPLE: The following margins are received at various steps along the chain of distribution
for a jar of pasta sauce that sells for a retail price of $5.00 (see Table 3.4).
What does it cost the manufacturer to produce a jar of pasta sauce? The retail selling price ($5.00),
multiplied by 1 less the retailer margin, will yield the wholesaler selling price. The wholesaler selling price can also be viewed as the cost to the retailer. The cost to the wholesaler (distributor selling
price) can be found by multiplying the wholesaler selling price by 1 less the wholesaler margin, and
so forth. Alternatively, one might follow the next procedure, using a channel member’s percentage
margin to calculate its dollar margin, and then subtracting that figure from the channel member’s
selling price to obtain its cost (see Table 3.5).
Thus, a jar of pasta that sells for $5.00 at retail actually costs the manufacturer 50 cents to make.
Table 3.4 Example—Pasta Sauce Distribution Margins
Distribution Stage
Margin
Manufacturer
50%
Distributor
50%
Wholesaler
33%
Retailer
40%
Table 3.5 Cost (Purchase Price) of Retailer
Stage
Margin %
$
Cost to Consumer
Retailer Margin
40%
$5.00
$2.00
Cost to Retailer
Wholesaler Margin
33%
$3.00
$1.00
Cost to Wholesaler
Distributor Margin
50%
$2.00
$1.00
Cost to Distributor
Manufacturer Margin
50%
$1.00
$0.50
Manufacturer’s Cost
$0.50
Chapter 3 Margins and Profits
81
The margins taken at multiple levels of a distribution process can have a dramatic
effect on the price paid by consumers. To work backward in analyzing these, many
people find it easier to convert markups to margins. Working forward does not require
this conversion.
EXAMPLE: To show that margins and markups are two sides of the same coin, let’s demonstrate that we can obtain the same sequence of prices by using the markup method here. Let’s look
at how the pasta sauce is marked up to arrive at a final consumer price of $5.00.
As noted previously, the manufacturer’s cost is $0.50. The manufacturer’s percentage markup is
100%. Thus, we can calculate its dollar markup as $0.50 * 100% = $0.50. Adding the manufacturer’s
markup to its cost, we arrive at its selling price: $0.50 (cost) + $0.50 (markup) = $1.00. The manufacturer sells the sauce for $1.00 to a distributor. The distributor applies a markup of 100%, taking the
price to $2.00, and sells the sauce to a wholesaler. The wholesaler applies a markup of 50% and sells
the sauce to a retailer for $3.00. Finally, the retailer applies a markup of 66.7% and sells the pasta
sauce to a consumer for $5.00. In Table 3.6, we track these markups to show the pasta sauce’s journey from a manufacturer’s cost of $0.50 to a retail price (consumer’s cost) of $5.00.
Table 3.6 Markups Along the Distribution Channel
Stage
Markup %
Manufacturer’s Cost
Manufacturer Markup
100%
$0.50
50%
Cost to Distributor
Distributor Markup
100%
$1.00
$1.00
50%
Cost to Wholesaler
Wholesaler Markup
50%
$2.00
$1.00
33.3%
Cost to Retailer
Retailer Markup
67%
$3.00
$2.00
40%
Cost to Consumer
82
MARKETING METRICS
$
Margin
$0.50
$5.00
Data Sources, Complications, and Cautions
The information needed to calculate channel margins is the same as for basic margins.
Complications arise, however, because of the layers involved. In this structure, the selling price for one layer in the chain becomes the cost to the next layer. This is clearly
visible in consumer goods industries, where there are often multiple levels of distribution between the manufacturer and the consumer, and each channel member requires
its own margin.
Cost and selling price depend on location within the chain. One must always ask,
“Whose cost is this?” and “Who sells at this price?” The process of “chaining” a
sequence of margins is not difficult. One need only clarify who sells to whom. In tracking this, it can help first to draw a horizontal line, labeling all the channel members
along the chain, with the manufacturer at the far left and the retailer on the right.
For example, if a beer exporter in Germany sells to an importer in the U.S., and that
importer sells to a distributor in Virginia, who sells the beer to a retailer, then four distinct selling prices and three channel margins will intervene between the exporter and
retail store customer. In this scenario, the exporter is the first supplier. The importer is
the first customer. To avoid confusion, we recommend mapping out the channel and
calculating margins, purchase prices, and selling prices at each level.
Throughout this section, we’ve assumed that all margins are “gross margins,” calculated
as selling price minus cost of goods sold. Of course, channel members will incur other
costs in the process of “adding value.” If a wholesaler pays his salespeople a commission
on sales, for example, that would be a cost of doing business. But it would not be a part
of the cost of goods sold, and so it is not factored into gross margin.
Related Metrics and Concepts
Hybrid (Mixed) Channel Margins
Hybrid Channel: The use of multiple distribution systems to reach the same
market. A company might approach consumers through stores, the Web, and
telemarketing, for example. Margins often differ among such channels. Hybrid
channels may also be known as mixed channels.
Increasingly, businesses “go to market” in more than one way. An insurance company,
for example, might sell policies through independent agents, toll-free telephone lines,
and the Web. Multiple channels often generate different channel margins and cause
a supplier to incur different support costs. As business migrates from one channel to
another, marketers must adjust pricing and support in economically sensible ways. To
make appropriate decisions, they must recognize the more profitable channels in their
mix and develop programs and strategies to fit these.
Chapter 3 Margins and Profits
83
When selling through multiple channels with different margins, it is important to perform analyses on the basis of weighted average channel margins, as opposed to a simple
average. Using a simple average can lead to confusion and poor decision-making.
As an example of the variations that can occur, let’s suppose that a company sells
10 units of its product through six channels. It sells five units through one channel at a 20% margin, and one unit through each of the other five channels at a 50%
margin. Calculating its average margin on a weighted basis, we arrive at the following
figure:
Percentage Margin (%) =
[(5 * 20%) + (5 * 50%)]
10
= 35%
By contrast, if we calculate the average margin among this firm’s six channels on a
simple basis, we arrive at a very different figure:
Percentage Margin (%) =
[(1 * 20%) + (5 * 50%)]
6
= 45%
This difference in margin could significantly blur management decision-making.
Average Margin
When assessing margins in dollar terms, use percentage of unit sales.
Average Margin ($) = [Percentage of Unit Sales through Channel 1 (%) * Margin
Earned in Channel 1 ($)] + [Percentage of Unit Sales
through Channel 2 (%) * Margin Earned in Channel 2 ($)]
+ Continued to Last Channel
When assessing margin in percentage terms, use percentage of dollar sales.
Average Margin (%) = [Percentage of Dollar Sales through Channel 1 (%) * Margin
Earned in Channel 1 (%)] + [Percentage of Dollar Sales
through Channel 2 (%) * Margin Earned in Channel 2 (%)]
+ Continued to Last Channel
EXAMPLE: Gael’s Glass sells through three channels: phone, online, and store. These channels
generate the following margins: 50%, 40%, and 30%, respectively. When Gael’s wife asks what his
average margin is, he initially calculates a simple margin and says it’s 40%. Gael’s wife investigates
further, however, and learns that her husband answered too quickly. Gael’s company sells a total of
10 units. It sells one unit by phone at a 50% margin, four units online at a 40% margin, and five units
84
MARKETING METRICS
in the store at a 30% margin. To determine the company’s average margin among these channels,
the margin in each must be weighted by its relative sales volume. On this basis, Gael’s wife calculates
the weighted average margin as follows:
Average Channel Margin = (Percentage of Unit Sales by Phone * Phone Channel Margin)
+ (Percentage of Unit Sales Online * Online Channel Margin)
+ (Percentage of Unit Sales through Store * Store Channel
Margin)
= (1/10 * 50%) + (4/10 * 40%) + (5/10 * 30%)
= 5% + 16% + 15%
Average Channel Margin = 36%
EXAMPLE:
results:
Sadetta, Inc. has two channels—online and retail—which generate the following
One customer orders online, paying $10 for one unit of goods that costs the company $5. This generates a 50% margin for Sadetta. A second customer shops at the store, buying two units of product
for $12 each. Each costs $9. Thus, Sadetta earns a 25% margin on these sales. Summarizing:
Online Margin (1) = 50%. Selling Price (1) = $10. Supplier Selling Price (1) = $5.
Store Margin (2) = 25%. Selling Price (2) = $12. Supplier Selling Price (2) = $9.
In this scenario, the relative weightings are easy to establish. In unit terms, Sadetta sells a total of
three units: one unit (33.3%) online, and two (66.6%) in the store. In dollar terms, Sadetta generates
a total of $34 in sales: $10 (29.4%) online, and $24 (70.6%) in the store.
Thus, Sadetta’s average unit margin ($) can be calculated as follows: The online channel generates
a $5.00 margin, while the store generates a $3.00 margin. The relative weightings are online 33.3%
and store 66.6%.
Average Unit Margin ($) = [Percentage Unit Sales Online (%) * Unit Margin Online ($)]
+ [Percentage Unit Sales in Store (%) * Unit Margin in Store ($)]
= 33.3% * $5.00 + 66.6% * $3.00
= $1.67 + $2.00
= $3.67
Sadetta’s average margin (%) can be calculated as follows: The online channel generates a 50% margin, while the store generates a 25% margin. The relative weightings are
online 29.4% and store 70.6%.
Chapter 3 Margins and Profits
85
Average Margin (%) = [Percentage Dollar Sales Online (%) * Margin Online (%)]
+ [Percentage Dollar Sales in Store (%) * Margin in Store (%)]
= 29.4% * 50% + 70.6% * 25%
= 14.70% + 17.65%
= 32.35%
Average margins can also be calculated directly from company totals. Sadetta, Inc. generated
a total gross margin of $11 by selling three units of product. Its average unit margin was thus
$11/3, or $3.67. Similarly, we can derive Sadetta’s average percentage margin by dividing its total
margin by its total revenue. This yields a result that matches our weighted previous calculations:
$11/$34 = 32.35%.
The same weighting process is needed to calculate average selling prices.
Average Selling Price ($) = [Percentage Unit Sales through Channel 1 (%)
* Selling Price in Channel 1 ($)] + [Percentage Unit Sales
through Channel 2 (%) * Selling Price in Channel 2 ($)]
+ Continued to [Percentage Unit Sales through the Last
Channel (%) * the Last Channel’s Selling Price ($)]
EXAMPLE: Continuing the previous example, we can see how Sadetta, Inc. calculates its average selling price.
Sadetta’s online customer pays $10 per item. Its store customer pays $12 per item. Weighting each
channel by unit sales, we can derive Sadetta’s average selling price as follows:
Average Selling Price ($) = [Percentage Unit Sales Online (%) * Selling Price Online ($)]
+ [Percentage Unit Sales in Store (%) * Selling Price in Store ($)]
= 33.3% * $10 + 66.7% * $12
= $3.33 + $8
= $11.33
The calculation of average supplier selling price is conceptually similar.
Average Supplier Selling Price ($) = [Percentage Unit Sales through Channel 1 (%)
* Supplier Selling Price in Channel 1 ($)] + [Percentage
Unit Sales through Channel 2 (%) * Supplier Selling
Price in Channel 2 ($)] + Continued to [Percentage
Unit Sales through the Last Channel (%) * the Last
Channel Supplier’s Selling Price ($)]
86
MARKETING METRICS
EXAMPLE: Now, let’s consider how Sadetta, Inc. calculates its average supplier selling price.
Sadetta’s online merchandise cost the company $5 per unit. Its in-store merchandise cost $9 per
unit. Thus:
Average Supplier Selling Price ($) = [Percentage Unit Sales Online (%) * Supplier Selling
Price Online ($)] + [Percentage Unit Sales through
Store (%) * Supplier Selling Price in Store ($)]
= 33.3% * $5 + 66.7% * $9
= $1.67 + $6 = $7.67
With all these pieces of the puzzle, we now have much greater insight into Sadetta, Inc.’s business
(see Table 3.7).
Table 3.7
Sadetta’s Channel Measures
Online
In Store
Selling Price (SP)
Supplier Selling Price (SSP)
Unit Margin ($)
Margin (%)
$10.00
$5.00
$5.00
50%
$12.00
$9.00
$3.00
25%
Units Sold
% Unit Sales
Dollar Sales
% Dollar Sales
Total Margin
1
33.3%
$10.00
29.4%
$5.00
2
66.7%
$24.00
70.6%
$6.00
Average Unit Margin ($)
Average Margin (%)
Average Selling Price
Average Supplier Selling Price
Average/Total
3
$34.00
$11.00
$3.67
32.4%
$11.33
$7.67
3.3 Average Price per Unit and Price per Statistical Unit
Average prices represent, quite simply, total sales revenue divided by total units sold.
Many products, however, are sold in multiple variants, such as bottle sizes. In these
cases, managers face a challenge: They must determine “comparable” units.
Chapter 3 Margins and Profits
87
Average prices can be calculated by weighting different unit selling prices by the percentage of unit sales (mix) for each product variant. If we use a standard, rather than
an actual mix of sizes and product varieties, the result is price per statistical unit.
Statistical units are also known as equivalent units.
Average Price per Unit ($) =
Revenue ($)
Units Sold (#)
or
= [Price of SKU 1 ($) * SKU 1 Percentage of Sales (%)]
+ [Price of SKU 2 ($) * SKU 2 Percentage of Sales (%)]
Price per Statistical Unit ($) = Total Price of a Bundle of SKUs Comprising
a Statistical Unit ($)
Unit Price per Statistical Unit ($) =
Price per Statistical Unit ($)
Total Units in the Bundle of SKUs
Comprising that Statistical Unit (#)
Average price per unit and prices per statistical unit are needed by marketers who
sell the same product in different packages, sizes, forms, or configurations at a
variety of different prices. As in analyses of different channels, these product and
price variations must be reflected accurately in overall average prices. If they are
not, marketers may lose sight of what is happening to prices and why. If the price
of each product variant remained unchanged, for example, but there was a shift
in the mix of volume sold, then the average price per unit would change, but the
price per statistical unit would not. Both of these metrics have value in identifying
market movements.
Purpose: To calculate meaningful average selling prices within a product line
that includes items of different sizes.
Many brands or product lines include multiple models, versions, flavors, colors, sizes,
or—more generally—stock keeping units (SKUs). Brita water filters, for example, are
sold in a number of SKUs. They are sold in single-filter packs, double-filter packs, and
special banded packs that may be restricted to club stores. They are sold on a standalone
basis and in combination with pitchers. These various packages and product forms may
be known as SKUs, models, items, and so on.
Stock Keeping Unit (SKU): A term used by retailers to identify individual items
that are carried or “stocked” within an assortment. This is the most detailed level
at which the inventory and sales of individual products are recorded.
88
MARKETING METRICS
Marketers often want to know both their own average prices and those of retailers. By
reckoning in terms of SKUs, they can calculate an average price per unit at any level in
the distribution chain. Two of the most useful of these averages are
1. A unit price average that includes all sales of all SKUs, expressed as an average
price per defined unit. In the water filter industry, for example, these might
include such figures as $2.23/filter, $0.03/filtered ounce, and so on.
2. A price per statistical unit that consists of a fixed bundle (number) of individual SKUs. This bundle is often constructed so as to reflect the actual mix of sales
of the various SKUs.
The average price per unit will change when there is a shift in the percentage of sales
represented by SKUs with different unit prices. It will also change when the prices of
the individual SKUs are modified. This contrasts with price per statistical unit, which,
by definition, has a fixed proportion of each SKU. Consequently, a price per statistical
unit will change only when there is a change in the price of one or more of the SKUs
included in it.
The information gleaned from a price per statistical unit can be helpful in considering
price movements within a market. Price per statistical unit, in combination with unit
price averages, provides insight into the degree to which the average prices in a market
are changing as a result of shifts in “mix”—proportions of sales generated by differently
priced SKUs—versus price changes for individual items. Alterations in mix—such as
a relative increase in the sale of larger versus smaller ice cream tubs at retail grocers,
for example—will affect average unit price, but not price per statistical unit. Pricing
changes in the SKUs that make up a statistical unit, however, will be reflected by a
change in the price of that statistical unit.
Construction
As with other marketing averages, average price per unit can be calculated either from
company totals or from the prices and shares of individual SKUs.
Average Price per Unit ($) =
Revenue ($)
Unit Sales (#)
or
= [Unit Price of SKU 1 ($) * SKU 1 Percentage of Sales (%)]
+ [Unit Price of SKU 2 ($) * SKU 2 Percentage of Sales (%)]
+ and so forth
The average price per unit depends on both unit prices and unit sales of individual
SKUs. The average price per unit can be driven upward by a rise in unit prices, or by
an increase in the unit shares of higher-priced SKUs, or by a combination of the two.
Chapter 3 Margins and Profits
89
An “average” price metric that is not sensitive to changes in SKU shares is the price per
statistical unit.
Price per Statistical Unit
Procter & Gamble and other companies face a challenge in monitoring prices for a wide
variety of product sizes, package types, and product formulations. Many consumer
brands have multiple SKUs and each SKU may have a different price. In these situations, how do marketers determine a brand’s overall price level in order to compare it
to competitive offerings or to track whether prices are rising or falling? One solution
is the “statistical unit,” also known as the “statistical case” or—in volumetric or weight
measures—the statistical liter or statistical ton. A statistical case of 288 ounces of liquid
detergent, for example, might be defined as comprising
Four 4-oz bottles = 16 oz
Twelve 12-oz bottles = 144 oz
Two 32-oz bottles = 64 oz
One 64-oz bottle = 64 oz
Note that the contents of this statistical case were carefully chosen so that it contains
the same number of ounces as a standard case of 24 12-ounce bottles. In this way, the
statistical case is comparable in size to a standard case. The advantage of a statistical
case is that its contents can approximate the mix of SKUs the company actually sells.
Whereas a statistical case of liquid detergent will be filled with whole bottles, in other
instances a statistical unit might contain fractions of certain packaging sizes in order for
its total contents to match a required volumetric or weight total.
Statistical units are composed of fixed proportions of different SKUs. These fixed proportions ensure that changes in the prices of the statistical unit reflect only changes in
the prices of the SKUs that comprise it.
The price of a statistical unit can be expressed either as a total price for the bundle of
SKUs comprising it, or in terms of that total price divided by the total volume of its
contents. The former might be called the “price per statistical unit”; the latter, the “unit
price per statistical unit.”
EXAMPLE: Carl’s Coffee Creamer (CCC) is sold in three sizes: a one-liter economy size, a halfliter “fridge-friendly” package, and a 0.05-liter single serving. Carl defines a 12-liter statistical case
of CCC as
Two units of the economy size = 2 liters (2 * 1.0 liter)
19 units of the fridge-friendly package = 9.5 liters (19 * 0.5 liter)
Ten single servings = 0.5 liter (10 * .05)
90
MARKETING METRICS
Prices for each size and the calculation of total price for the statistical unit are shown in the following table:
Number
in Statistical
Case
Liters in
Statistical
Case
Total
Price
SKU Names
Size
Price of
Item
Economy
1 Liter
$8.00
2
2.0
$16.00
Fridge-Friendly
0.5 Liter
$6.00
19
9.5
$114.00
Single Serving
0.05 Liter
$1.00
10
0.5
$10.00
12
$140.00
TOTAL
Thus, the total price of the 12-liter statistical case of CCC is $140. The per-liter price within the
statistical case is $11.67.
Note that the $140 price of the statistical case is higher than the $96 price of a case of 12 economy
packs. This higher price reflects the fact that smaller packages of CCC command a higher price per
liter. If the proportions of the SKUs in the statistical case exactly match the actual proportions
sold, then the per-liter price of the statistical case will match the per-liter price of the actual liters
sold.
EXAMPLE: Carl sells 10,000 one-liter economy packs of CCC, 80,000 fridge-friendly half
liters, and 40,000 single servings. What was his average price per liter?
Average Price per Unit ($) =
Revenue ($)
Unit Sales (#)
($8 * 10k + $6 * 80k + $1 * 40k)
=
=
(1 * 10k + 0.5 * 80k + 0.05 * 40k)
$600k
52k
= $11.54
Note that Carl’s average price per liter, at $11.54, is less than the per-liter price in his statistical case.
The reason is straightforward: Whereas fridge-friendly packs outnumber economy packs by almost
ten to one in the statistical case, the actual sales ratio of these SKUs was only eight to one. Similarly,
whereas the ratio of single-serving items to economy items in the statistical case is five to one, their
actual sales ratio was only four to one. Carl’s company sold a smaller percentage of the higher (per
liter) priced items than was represented in its statistical case. Consequently, its actual average price
per liter was less than the per-liter price within its statistical unit.
Chapter 3 Margins and Profits
91
In the following table, we illustrate the calculation of the average price per unit as the
weighted average of the unit prices and unit shares of the three SKUs of Carl’s Coffee
Creamer. Unit prices and unit (per-liter) shares are provided.
SKU Name
Size
Price
SKUs
Sold
Units Sold
(Liters)
Unit Price
(per Liter)
Unit
Share
Economy
1 Liter
$8
10k
10k
$8
19.23%
Fridge-Friendly
0.5 Liter
$6
80k
40k
$12
76.92%
Single Serving
0.05 Liter
$1
40k
2k
$20
3.85%
130k
52k
TOTAL
100%
On this basis, the average price per unit ($) = ($8 * 0.1923) + ($12 * 0.7692) + ($20 * 0.0385)
= $11.54.
Data Sources, Complications, and Cautions
With complex and changing product lines, and with different selling prices charged
by different retailers, marketers need to understand a number of methodologies for
calculating average prices. Merely determining how many units of a product are sold,
and at what price, throughout the market is a major challenge. As a standard method of
tracking prices, marketers use statistical units, which are based on constant proportions
of sales of different SKUs in a product line.
Typically, the proportions of SKUs in a statistical unit correspond—at least
approximately—to historical market sales. Sales patterns can change, however. In consequence, these proportions need to be monitored carefully in evolving markets and
changing product lines.
Calculating a meaningful average price is complicated by the need to differentiate
between changes in sales mix and changes in the prices of statistical units. In some
industries, it is difficult to construct appropriate units for analyzing price and sales data.
In the chemical industry, for example, an herbicide might be sold in a variety of different sizes, applicators, and concentration levels. When we factor in the complexity of
different prices and different assortments offered by competing retail outlets, calculating and tracking average prices becomes a non-trivial exercise.
Similar challenges arise in estimating inflation. Economists calculate inflation by using
a basket of goods. Their estimates might vary considerably, depending on the goods
included. It is also difficult to capture quality improvements in inflation figures. Is a
2015 car, for example, truly comparable to a car built 30 years earlier?
In evaluating price increases, marketers are advised to bear in mind that a consumer who shops for large quantities at discount stores may view such increases very
92
MARKETING METRICS
differently from a pensioner who buys small quantities at local stores. Establishing a
“standard” basket for such different consumers requires astute judgment. In seeking to
summarize the aggregate of such price increases throughout an economy, economists
may view inflation as, in effect, a statistical unit price measure for that economy.
3.4 Variable Costs and Fixed Costs
Variable costs can be aggregated into a “total” or expressed on a “per-unit” basis.
Fixed costs, by definition, do not change with the number of units sold or produced.
Variable costs are assumed to be relatively constant on a per-unit basis. Total variable costs increase directly and predictably with unit sales volume. Fixed costs, on
the other hand, do not change as a direct result of short-term unit sales increases or
decreases.
Total Costs ($) = Fixed Costs ($) + Total Variable Costs ($)
Total Variable Costs ($) = Unit Volume (#) * Variable Cost per Unit ($)
Marketers need to have an idea of how costs divide between variable and fixed. This
distinction is crucial in forecasting the earnings generated by various changes in unit
sales and thus the financial impact of proposed marketing campaigns. It is also fundamental to an understanding of price and volume trade-offs.
Purpose: To understand how costs change with volume.
At first glance, this appears to be an easy subject to master. If a marketing campaign will
generate 10,000 units of additional sales, we need only know how much it will cost to
supply that additional volume.
The problem, of course, is that no one really knows how changes in quantity will affect a
firm’s total costs—in part because the workings of a firm can be so complex. Companies
simply can’t afford to employ armies of accountants to answer every possible expense
question precisely. Instead, we often use a simple model of cost behavior that is good
enough for most purposes.
Construction
The standard linear equation, Y = mX + b, helps explain the relationship between total
costs and unit volume. In this application, Y will represent a company’s total cost, m
will be its variable cost per unit, X will represent the quantity of products sold (or produced), and b will represent the fixed cost (see Figure 3.3).
Total Cost ($) = Variable Cost per Unit ($) * Quantity (#) + Fixed Cost ($)
Chapter 3 Margins and Profits
93
Total Cost $ (Y)
Fixed and Variable Costs
Effect of adding 1 unit variable cost (m)
}
Fixed Cost (b)
Quantity (X)
Figure 3.3 Fixed and Variable Costs
On this basis, to determine a company’s total cost for any given quantity of products,
we need only multiply its variable cost per unit by that quantity and add its fixed cost.
To communicate fully the implications of fixed costs and variable costs, it may help to
separate this graph into two parts (see Figure 3.4).
By definition, fixed costs remain the same, regardless of volume. Consequently, they
are represented by a horizontal line across the graph in Figure 3.4. Fixed costs do not
increase vertically—that is, they do not add to the total cost—as quantity rises.
The result of multiplying variable cost per unit by quantity is often called the total variable cost. Variable costs differ from fixed costs in that, when there is no production,
their total is zero. Their total increases in a steadily rising line, however, as quantity
increases.
We can represent this model of cost behavior in a simple equation.
Total Cost ($) = Total Variable Cost ($) + Fixed Cost ($)
94
MARKETING METRICS
Total Cost $ (Y)
Fixed Costs
Fixed Cost (b)
Quantity (X)
Total Cost $ (Y)
Variable Costs
Total Variable Cost (m*X)
Quantity (X)
Figure 3.4
Total Cost Consists of Fixed and Variable Costs
Chapter 3 Margins and Profits
95
To use this model, of course, we must place each of a firm’s costs into one or the other
of these two categories. If an expense does not change with volume (rent, for example),
then it is part of fixed costs and will remain the same, regardless of how many units
the firm produces or sells. If a cost does change with volume (sales commissions, for
example), then it is a variable cost.
Total Variable Costs ($) = Unit Volume (#) * Variable Cost per Unit ($)
Total Cost per Unit: It is also possible to express the total cost for a given quantity on
a per-unit basis. The result might be called total cost per unit, unit total cost, average
cost, full cost, or even fully loaded cost. For our simple linear cost model, the total cost
per unit can be calculated in either of two ways. The most obvious would be to divide
the total cost by the number of units.
Total Cost per Unit ($) =
Total Cost ($)
Quantity (#)
This can be plotted graphically, and it tells an interesting tale (see Figure 3.5). As the
quantity rises, the total cost per unit (average cost per unit) declines. The shape of this
Unit Total Cost ($)
Effects of Fixed and Variable Costs on Cost per Unit
Quantity
Figure 3.5
96
Total Cost per Unit Falls with Volume (Typical Assumptions)
MARKETING METRICS
curve will vary among firms with different cost structures, but wherever there are both
fixed and variable costs, the total cost per unit will decline as fixed costs are spread
across an increasing quantity of units.
The apportionment of fixed costs across units produced leads us to another common
formula for the total cost per unit.
Total Cost per Unit ($) = Variable Cost per Unit ($) + [Fixed Cost ($)/Quantity (#)]
As the quantity increases—that is, as fixed costs are spread over an increasing number
of units—the total cost per unit declines in a non-linear way.3
EXAMPLE: As a company’s unit sales increase, its fixed costs hold steady at $500. The variable
cost per unit remains constant at $10 per unit. Total variable costs increase with each unit sold. The
total cost per unit (also known as average total cost) decreases as incremental units are sold and as
fixed costs are spread across this rising quantity. Eventually, as more and more units are produced
and sold, the company’s total cost per unit approaches its variable cost per unit (see Table 3.8).
Table 3.8 Fixed and Variable Costs at Increasing Volume Levels
Units Sold
Fixed Costs
Variable Costs
Total Costs
1
$500
$10
$510
10
$500
$100
$600
100
$500
$1,000
$1,500
1,000
$500
$10,000
$10,500
Total Cost per Unit
Variable Cost per Unit
$510.00
$10
$60.00
$10
$15.00
$10
$10.50
$10
In summary, the simplest model of cost behavior is to assume total costs increase linearly with quantity supplied. Total costs are composed of fixed and variable costs. Total
cost per unit decreases in a non-linear way with rising quantity supplied.
Data Sources, Complications, and Cautions
Total cost is typically assumed to be a linear function of quantity supplied. That is, the
graph of total cost versus quantity will be a straight line. Because some costs are fixed,
total cost starts at a level above zero, even when no units are produced. This is because
fixed costs include such expenses as factory rent and salaries for full-time employees,
Chapter 3 Margins and Profits
97
which must be paid regardless of whether any goods are produced and sold. Total
variable costs, by contrast, rise and fall with quantity. Within our model, however, variable cost per unit is assumed to hold constant—at $10 per unit for example—regardless
of whether one unit or 1,000 units are produced. This is a useful model. In using it,
however, marketers must recognize that it fails to account for certain complexities.
The linear cost model does not fit every situation: Quantity discounts, expectations of
future process improvements, and capacity limitations, for example, introduce dynamics
that will limit the usefulness of the fundamental linear cost equation: Total Cost = Fixed
Cost + Variable Cost per Unit * Quantity. Even the notion that quantity determines the
total cost can be questioned. Although firms pay for inputs, such as raw materials and
labor, marketers want to know the cost of the firm’s outputs, that is, finished goods sold.
This distinction is clear in theory. In practice, however, it can be difficult to uncover the
precise relationship between a quantity of outputs and the total cost of the wide array of
inputs that go into it.
The classification of costs as fixed or variable depends on context: Even though the
linear model may not work in all situations, it does provide a reasonable approximation
for cost behavior in many contexts. Some marketers have trouble, however, with the
fact that certain costs can be considered fixed in some contexts and variable in others.
In general, for shorter time frames and modest changes in quantity, many costs are
fixed. For longer time frames and larger changes in quantity, most costs are variable.
Let’s consider rent, for example. Small changes in quantity do not require a change in
workspace or business location. In such cases, rent should be regarded as a fixed cost. A
major change in quantity, however, would require more or less workspace. Rent, therefore, would become variable over that range of quantity.
Don’t confuse Total Cost per Unit with Variable Cost per Unit: In our linear cost
equation, the variable cost per unit is the amount by which total costs increase if the
firm increases its quantity by one unit. This number should not be confused with the
total cost per unit, calculated as Variable Cost per Unit + (Fixed Cost/Quantity). If a
firm has fixed costs, then its total cost per unit will always be greater than the variable
cost per unit. Total cost per unit represents the firm’s average cost per unit at the current quantity—and only at the current quantity. Do not make the mistake of thinking
of total cost per unit as a figure that applies to changing quantities. Total cost per unit
only applies at the volume at which it was calculated.
A related misunderstanding may arise at times from the fact that total cost per unit generally decreases with rising quantity. Some marketers use this fact to argue for aggressively increasing quantity in order to “bring our costs down” and improve profitability.
Total cost, by contrast with total cost per unit, almost always increases with quantity.
Only with certain quantity discounts or rebates that “kick in” when target volumes are
reached can total cost decrease as volume increases.
98
MARKETING METRICS
3.5 Marketing Spending—Total, Fixed, and Variable
To predict how selling costs change with sales, a firm must distinguish between fixed
selling costs and variable selling costs.
Total Selling (Marketing) Costs ($) = Total Fixed Selling Costs ($)
+ Total Variable Selling Costs ($)
Total Variable Selling Costs ($) = Revenue ($) * Variable Selling Cost (%)
Recognizing the difference between fixed and variable selling costs can help firms
account for the relative risks associated with alternative sales strategies. In general,
strategies that incur variable selling costs are less risky because variable selling costs
will remain lower in the event that sales fail to meet expectations.
Purpose: To forecast marketing spending and assess budgeting risk.
Marketing Spending: Total expenditure on marketing activities. This typically
includes advertising and non-price promotion. It sometimes includes sales force
spending and may also include price promotions.
Marketing costs are often a major part of a firm’s overall discretionary expenditures.
As such, they are important determinants of short-term profits. Of course, marketing
and selling budgets can also be viewed as investments in acquiring and maintaining
customers. From either perspective, however, it is useful to distinguish between fixed
marketing costs and variable marketing costs. That is, managers must recognize which
marketing costs will hold steady, and which will change with sales. Generally, this classification will require a “line-item by line-item” review of the entire marketing budget.
In prior sections, we have viewed total variable costs as expenses that vary with unit
sales volume. With respect to selling costs, we’ll need a slightly different conception.
Rather than varying with unit sales, total variable selling costs are more likely to vary
directly with the monetary value of the units sold—that is, with revenue. Thus, it is
more likely that variable selling costs will be expressed as a percentage of revenue,
rather than a certain monetary amount per unit.
The classification of selling costs as fixed or variable will depend on an organization’s
structure and on the specific decisions of management. A number of items, however,
typically fall into one category or the other—with the proviso that their status as fixed
or variable can be time-specific. In the long run, all costs eventually become variable.
Over typical planning periods of a quarter or a year, fixed marketing costs might include
■
Sales force salaries and support.
■
Major advertising campaigns, including production costs.
Chapter 3 Margins and Profits
99
■
■
■
Marketing staff.
Sales promotion material, such as point-of-purchase sales aids, coupon production, and distribution costs.
Cooperative advertising allowances based on prior-period sales.
Variable marketing costs might include
■
Sales commissions paid to sales force, brokers, or manufacturer representatives.
■
Sales bonuses contingent on reaching sales goals.
■
Off-invoice and performance allowances to trade, which are tied to current
volume.
■
Early payment terms (if included in sales promotion budgets).
■
Coupon face-value payments and rebates, including processing fees.
■
Bill-backs for local campaigns, which are conducted by retailers but reimbursed
by national brand and cooperative advertising allowances, based on current
period sales.
Marketers often don’t consider their budgets in fixed and variable terms, but they can
derive at least two benefits by doing so.
First, if marketing spending is in fact variable, then budgeting in this way is more accurate.
Some marketers budget a fixed amount and then face an end-of-period discrepancy
or “variance” if sales miss their declared targets. By contrast, a flexible budget—that is,
one that takes account of its genuinely variable components—will reflect actual results,
regardless of where sales end up.
Second, the short-term risks associated with fixed marketing costs are greater than
those associated with variable marketing costs. If marketers expect revenues to be
sensitive to factors outside their control—such as competitive actions or production
shortages—they can reduce risk by including more variable and less fixed spending in
their budgets.
A classic decision that hinges on fixed marketing costs versus variable marketing costs
is the choice between engaging third-party contract sales representatives versus an
in-house sales force. Hiring a salaried—or predominantly salaried—sales force entails
more risk than the alternative because salaries must be paid even if the firm fails to
achieve its revenue targets. By contrast, when a firm uses third-party brokers to sell its
goods on commission, its selling costs decline when sales targets are not met.
Construction
Total Selling (Marketing) Costs ($) = Total Fixed Selling Costs ($)
+ Total Variable Selling Costs ($)
Total Variable Selling Costs ($) = Revenue ($) * Variable Selling Cost (%)
100
MARKETING METRICS
Commissioned Sales Costs: Sales commissions represent one example of selling costs
that vary in proportion to revenue. Consequently, any sales commissions should be
included in variable selling costs.
EXAMPLE: Henry’s Catsup spends $10 million a year to maintain a sales force that calls on grocery chains and wholesalers. A broker offers to perform the same selling tasks for a 5% commission.
At $100 million in revenue,
Total Variable Selling Cost = $100 million * 5% = $5 million
At $200 million in revenue,
Total Variable Selling Cost = $200 million * 5% = $10 million
At $300 million in revenue,
Total Variable Selling Cost = $300 million * 5% = $15 million
If revenues run less than $200 million, the broker will cost less than the in-house sales force. At $200
million in revenue, the broker will cost the same as the sales force. At revenue levels greater than
$200 million, the broker will cost more.
Of course, the transition from a salaried sales staff to a broker may itself cause a change in revenues.
Calculating the revenue level at which selling costs are equal is only a starting point for analysis. But
it is an important first step in understanding the trade-offs.
There are many types of variable selling costs. For example, selling costs could be based
upon a complicated formula, specified in a firm’s contracts with its brokers and dealers.
Selling costs might include incentives to local dealers, which are tied to the achievement
of specific sales targets. They might include promises to reimburse retailers for spending on cooperative advertising. By contrast, payments to a Web site for a fixed number
of impressions or click-throughs, in a contract that calls for specific dollar compensation, would more likely be classified as fixed costs. On the other hand, payments for
conversions (sales) would be classified as variable marketing costs.
EXAMPLE: A small manufacturer of a regional food delicacy must select a budget for a television advertising campaign that it plans to launch. Under one plan, it might pay to create a commercial and air it in a certain number of time slots. Its spending level would thus be fixed. It would be
selected ahead of time and would not vary with the results of the campaign.
Under an alternative plan, the company could produce the advertisement—still a fixed cost—but
ask retailers to air it in their local markets and pay the required media fees to television stations
as part of a cooperative advertising arrangement. In return for paying the media fees, local stores
would receive a discount (a bill-back) on every unit of the company’s product that they sell.
Chapter 3 Margins and Profits
101
Under the latter plan, the product discount would be a variable cost, as its total amount would
depend on the number of units sold. By undertaking such a cooperative advertising campaign,
the manufacturer would make its marketing budget a mix of fixed and variable costs. Is such
cooperative advertising a good idea? To decide this, the company must determine its expected sales
under both arrangements, as well as the consequent economics and its tolerance for risk.
Data Sources, Complications, and Cautions
Fixed costs are often easier to measure than variable costs. Typically, fixed costs might
be assembled from payroll records, lease documents, or financial records. For variable
costs, it is necessary to measure the rate at which they increase as a function of activity
level. Although variable selling costs often represent a predefined percentage of revenue, they may alternatively vary with the number of units sold (as in a dollar-per-case
discount). An additional complication arises if some variable selling costs apply to only
a portion of total sales. This can happen, for example, when some dealers qualify for
cash discounts or full-truckload rates and some do not.
In a further complication, some expenses may appear to be fixed when they are actually
stepped. That is, they are fixed to a point, but they trigger further expenditures beyond
that point. For example, a firm may contract with an advertising agency for up to three
campaigns per year. If it decides to buy more than three campaigns, it would incur an
incremental cost. Typically, stepped costs can be treated as fixed—provided that the
boundaries of analysis are well understood.
Stepped payments can be difficult to model. Rebates for customers whose purchases
exceed a certain level, or bonuses for salespeople who exceed quota, can be challenging
functions to describe. Creativity is important in designing marketing discounts. But this
creativity can be difficult to reflect in a framework of fixed and variable costs.
In developing their marketing budgets, firms must decide which costs to expense in the
current period and which to amortize over several periods. The latter course is appropriate for expenditures that are correctly viewed as investments. One example of such an
investment would be a special allowance for financing receivables from new distributors.
Rather than adding such an allowance to the current period’s budget, it would be better
viewed as a marketing item that increases the firm’s investment in working capital. By
contrast, advertising that is projected to generate long-term impact may be loosely called
an investment, but it would be better treated as a marketing expense. Although there
may be a valid theoretical case for amortizing advertising, that discussion is beyond the
scope of this book.
102
MARKETING METRICS
Related Metrics and Concepts
Levels of marketing spending are often used to compare companies and to demonstrate
how heavily they “invest” in this area. For this purpose, marketing spending is generally
viewed as a percentage of sales.
Marketing As a Percentage of Sales: The level of marketing spending as a fraction of
sales. This figure provides an indication of how heavily a company is marketing. The
appropriate level for this figure varies among products, strategies, and markets.
Marketing As a Percentage of Sales (%) =
Marketing Spending ($)
Revenue ($)
Variants on this metric are used to examine components of marketing in comparison
with sales. Examples include trade promotion as a percentage of sales, or sales force as
a percentage of sales. One particularly common example is:
Advertising As a Percentage of Sales: Advertising expenditures as a fraction of
sales. Generally, this is a subset of marketing as a percentage of sales.
Before using such metrics, marketers are advised to determine whether certain marketing costs have already been subtracted in the calculation of sales revenue. Trade allowances, for example, are often deducted from “gross sales” to calculate “net sales.”
Slotting Allowances: These are a particular form of selling costs encountered when
new items are introduced to retailers or distributors. Essentially, they represent a
charge made by retailers for making a “slot” available for a new item in their stores and
warehouses. This charge may take the form of a one-time cash payment, free goods, or
a special discount. The exact terms of the slotting allowance will determine whether it
constitutes a fixed or a variable selling cost, or a mix of the two.
3.6 Break-Even Analysis and Contribution Analysis
The break-even level represents the sales amount—in either unit or revenue terms—
that is required to cover total costs (both fixed and variable). Profit at break-even is
zero. Break-even is only possible if a firm’s prices are higher than its variable costs
per unit. If so, then each unit of product sold will generate some “contribution”
toward covering fixed costs. The difference between price per unit and variable cost
per unit is defined as Contribution per Unit.
Contribution per Unit ($) = Selling Price per Unit ($) – Variable Cost per Unit ($)4
Contribution Margin (%) =
Contribution per Unit ($)
Selling Price per Unit ($)
Chapter 3 Margins and Profits
103
Break-Even Volume (#) =
Fixed Costs ($)
Contribution per Unit ($)
Break-Even Revenue ($) = Break-Even Volume (Units) (#) * Price per Unit ($)
or
=
Fixed Costs ($)
Contribution Margin (%)
Break-even analysis is the Swiss Army knife of marketing economics. It is useful in a
variety of situations and is often used to evaluate the likely profitability of marketing
actions that affect fixed costs, prices, or variable costs per unit. Break-even is often
derived in a “back-of-the-envelope” calculation that determines whether a more
detailed analysis is warranted.
Purpose: To provide a rough indicator of the earnings impact
of a marketing activity.
The break-even point for any business activity is defined as the level of sales at which
neither a profit nor a loss is made on that activity—that is, where Total Revenues = Total
Costs. Provided that a company sells its goods at a price per unit that is greater than its
variable cost per unit, the sale of each unit will make a “contribution” toward covering
some portion of fixed costs. That contribution can be calculated as the difference between
price per unit (revenue) and variable cost per unit. On this basis, break-even constitutes
the minimum level of sales at which total contribution fully covers fixed costs.
Construction
To determine the break-even point for a business program, one must first calculate the
fixed costs of engaging in that program. For this purpose, managers do not need to estimate projected volumes. Fixed costs are constant, regardless of activity level. Managers
do, however, need to calculate the difference between revenue per unit and variable
costs per unit. This difference represents contribution per unit ($). Contribution rates
can also be expressed as a percentage of selling price.
EXAMPLE: Apprentice Mousetraps wants to know how many units of its “Magic Mouse
Trapper” it must sell to break even. The product sells for $20. It costs $5 per unit to make. The
company’s fixed costs are $30,000. Break-even will be reached when total contribution equals fixed
costs.
104
MARKETING METRICS
Break-Even Volume =
Fixed Costs
Contribution per Unit
Contribution per Unit = Sale Price per Unit – Variable Cost per Unit
= $20 – $5 = $15
Break-Even Volume =
$30,000
= 2,000 mousetraps
$15
This dynamic can be summarized in a graph that shows fixed costs, variable costs, total
costs, and total revenue (see Figure 3.6). Below the break-even point, total costs exceed
total revenue, creating a loss. Above the break-even point, a company generates profits.
Break-Even: Break-even occurs when the total contribution equals the fixed
costs. Profits and losses at this point equal zero.
One of the key building blocks of break-even analysis is the concept of contribution.
Contribution represents the portion of sales revenue that is not consumed by variable
costs and so contributes to the coverage of fixed costs.
Contribution per Unit ($) = Selling Price per Unit ($) – Variable Cost per Unit ($)
Profit is earned when
Total Revenue exceeds
Total Cost
The Break-Even Point
$
Break-Even
Below Break-Even,
Total Costs exceed
Total Revenue
Total Costs
Total Revenue
Fixed Costs
Total Variable
Costs
Units
Figure 3.6
At Break-Even, Total Costs = Total Revenues
Chapter 3 Margins and Profits
105
Contribution can also be expressed in percentage terms, quantifying the fraction of the
sales price that contributes toward covering fixed costs. This percentage is often called
the contribution margin.
Contribution Margin (%) =
Contribution per Unit ($)
Selling Price per Unit ($)
Formulas for total contribution include the following:
Total Contribution ($) = Units Sold (#) * Contribution per Unit ($)
Total Contribution ($) = Total Revenues ($) – Total Variable Costs ($)
As previously noted,
Total Variable Costs = Variable Costs per Unit * Units Sold
Total Revenues = Selling Price per Unit * Units Sold
Break-Even Volume: The number of units that must be sold to cover fixed costs.
Break-Even Volume (#) =
Fixed Costs ($)
Contribution per Unit ($)
Break-even will occur when an enterprise sells enough units to cover its fixed costs. If
the fixed costs are $10 and the contribution per unit is $2, then a firm must sell five
units to break even.
Break-Even Revenue: The level of dollar sales required to break even.
Break-Even Revenue ($) = Break-Even Volume (Units) (#) * Price per Unit ($)
This formula is the simple conversion of volume in units to the revenues generated by
that volume.
EXAMPLE: Apprentice Mousetraps wants to know how many dollars’ worth of its “Deluxe
Mighty Mouse Trapper” it must sell to break even. The product sells for $40 per unit. It costs $10
per unit to make. The company’s fixed costs are $30,000.
With fixed costs of $30,000, and a contribution per unit of $30, Apprentice must sell
$30,000/$30 = 1,000 deluxe mousetraps to break even. At $40 per trap, this corresponds to revenues of 1,000 * $40 = $40,000.
Break-Even Revenue ($) = Break-Even Volume (#) * Price per Unit ($)
= 1,000 * $40 = $40,000
Break-even in dollar terms can also be calculated by dividing fixed costs by the fraction of the selling
price that represents contribution.
106
MARKETING METRICS
Break-Even Revenue =
=
=
Fixed Costs
(Selling Price – Variable Costs)/Selling Price
$30,000
($40 – $10)/$40
$30,000
75%
= $40,000
Break-Even on Incremental Investment
Break-even on incremental investment is a common form of break-even analysis. It
examines the additional investment needed to pursue a marketing plan, and it calculates the additional sales required to cover that expenditure. Any costs or revenues
that would have occurred regardless of the investment decision are excluded from this
analysis.
EXAMPLE: John’s Clothing Store employs three salespeople. It generates annual sales of
$1 million and an average contribution margin of 30%. Rent is $50,000. Each sales person costs
$50,000 per year in salary and benefits. How much would sales have to increase for John to break
even on hiring an additional salesperson?
If the additional “investment” in a salesperson is $50,000, then break-even on the new hire will be
reached when sales increase by $50,000/ 30%, or $166,666.67.
Data Sources, Complications, and Cautions
To calculate a break-even sales level, one must know the revenues per unit, the variable
costs per unit, and the fixed costs. To establish these figures, one must classify all costs
as either fixed (those that do not change with volume) or variable (those that increase
linearly with volume).
The time scale of the analysis can influence this classification. Indeed, one’s managerial
intent can be reflected in the classification. (Will the company fire employees and sublet
factory space if sales turn down?) As a general rule, all costs become variable in the long
term. Firms generally view rent, for example, as a fixed cost. But in the long term, even
rent becomes variable as a company may move into larger quarters when sales grow
beyond a certain point.
Before agonizing over these judgments, managers are urged to remember that the
most useful application of the break-even exercise is to make a rough judgment about
whether more detailed analyses are likely to be worth the effort. The break-even calculation enables managers to judge various options and proposals quickly. It is not,
Chapter 3 Margins and Profits
107
however, a substitute for more detailed analyses, including projections of target profits
(Section 3.7), risk, and the time value of money (Sections 5.3 and 10.4).
Related Metrics and Concepts
Payback Period: The period of time required to recoup the funds expended in an
investment. The payback period is the time required for an investment to reach breakeven (see previous sections).
3.7 Profit-Based Sales Targets
In launching a program, managers often start with an idea of the dollar profit they
desire and ask what sales levels will be required to reach it. Target volume (#) is the
unit sales quantity required to meet an earnings goal. Target revenue ($) is the corresponding figure for dollar sales. Both of these metrics can be viewed as extensions
of break-even analysis.
Target Volume (#) =
[Fixed Costs ($) + Target Profits ($)]
Contribution per Unit ($)
Target Revenue ($) = Target Volume (#) * Selling Price per Unit ($)
or
=
[Fixed Costs ($) + Target Profits ($)]
Contribution Margin (%)
Increasingly, marketers are expected to generate volumes that meet the target profits
of their firm. This will often require them to revise sales targets as prices and costs
change.
Purpose: To ensure that marketing and sales objectives mesh
with profit targets.
In the previous section, we explored the concept of break-even, the point at which
a company sells enough to cover its fixed costs. In target volume and target revenue
calculations, managers take the next step. They determine the level of unit sales or revenues needed not only to cover a firm’s costs but also to attain its profit targets.
108
MARKETING METRICS
Construction
Target Volume: The volume of sales necessary to generate the profits specified in
a company’s plans.
The formula for target volume will be familiar to those who have performed break-even
analysis. The only change is to add the required profit target to the fixed costs. From
another perspective, the break-even volume equation can be viewed as a special case
of the general target volume calculation—one in which the profit target is zero, and a
company seeks only to cover its fixed costs. In target volume calculations, the company
broadens this objective to solve for a desired profit.
Target Volume (#) =
Fixed Costs ($) + Target Profits ($)
Contribution per Unit ($)
EXAMPLE: Mohan, an artist, wants to know how many caricatures he must sell to realize a
yearly profit objective of $30,000. Each caricature sells for $20 and costs $5 in materials to make.
The fixed costs for Mohan’s studio are $30,000 per year:
Target Volume =
=
Fixed Costs + Target Profits
Sales Price – Variable Costs
$30,000 + $30,000
$20 – $5
= 4,000 caricatures per year
It is quite simple to convert unit target volume to target revenues. One need only multiply the
volume figure by an item’s price per unit. Continuing the example of Mohan’s studio,
Target Revenue ($) = Target Volume (#) * Selling Price ($)
= 4,000 * $20 = $80,000
Alternatively, we can use a second formula:
Target Revenue =
=
=
Fixed Costs ($) + Target Profits ($)
Contribution Margin (%)
$30,000 + $30,00
$15/$20
$60,000
0.75
= $80,000
Chapter 3 Margins and Profits
109
Data Sources, Complications, and Cautions
The information needed to perform a target volume calculation is essentially the same
as that required for break-even analysis—fixed costs, selling price, and variable costs. Of
course, before determining target volume, one must also set a profit target.
The major assumption here is the same as in break-even analysis: Costs are linear with
respect to unit volume over the range explored in the calculation.
Related Metrics and Concepts
Target Volumes Not Based on Target Profit: In this section, we have assumed that a
firm starts with a profit target and seeks to determine the volume required to meet it.
In certain instances, however, a firm might set a volume target for reasons other than
short-term profit. For example, firms sometimes adopt top-line growth as a goal. Please
do not confuse this use of target volume with the profit-based target volumes calculated
in this section.
Returns and Targets: Companies often set hurdle rates for return on sales and return
on investment and require that projections achieve these before any plan can be
approved. Given these targets, we can calculate the sales volume required for the necessary return. (See Section 10.2 for more details.)
EXAMPLE: Niesha runs business development at Gird, a company that has established a
return on sales target of 15%. That is, Gird requires that all programs generate profits equivalent to
15% of sales revenues. Niesha is evaluating a program that will add $1,000,000 to fixed costs. Under
this program, each unit of product will be sold for $100 and will generate a contribution margin of
25%. To reach break-even on this program, Gird must sell $1,000,000/$25 = 40,000 units of product. How much must Gird sell to reach its target return on sales (ROS) of 15%?
To determine the revenue level required to achieve a 15% ROS, Niesha can use either a spreadsheet
model and trial and error, or the following formula:
Fixed Costs ($)
Target Revenue =
Contribution Margin (%) – Target ROS (%)
$1,000,000
=
=
0.25 – 0.15
$1,000,000
0.1
= $10,000,000
Thus, Gird will achieve its 15% ROS target if it generates $10,000,000 in sales. At a selling price of
$100 per unit, this is equivalent to unit sales of 100,000.
110
MARKETING METRICS
4
PRODUCT AND PORTFOLIO
MANAGEMENT
Introduction
Key concepts covered in this chapter:
Trial, Repeat, Penetration,
and Volume Projections
Conjoint Utilities and Consumer
Preference
Growth: Percentage and CAGR
Segmentation and Conjoint
Utilities
Cannibalization Rate and Fair
Share Draw Rate
Brand Equity Metrics
Conjoint Utilities and Volume
Projection
Effective marketing comes from customer knowledge and an understanding of how a
product fits customers’ needs. In this chapter, we’ll describe metrics used in product
strategy and planning. These metrics address the following questions: What volumes
can marketers expect from a new product? How will sales of existing products be
affected by the launch of a new offering? Is brand equity increasing or decreasing?
What do customers really want, and what are they willing to sacrifice to obtain it?
We’ll start with a section on trial and repeat rates, explaining how these metrics are
determined and how they’re used to generate sales forecasts for new products. Because
forecasts involve growth projections, we’ll then discuss the difference between yearon-year growth and compound annual growth rates (CAGR). Because growth of one
product sometimes comes at the expense of an existing product line, it is important
111
to understand cannibalization metrics. These reflect the impact of new products on a
portfolio of existing products.
Next, we’ll cover selected metrics associated with brand equity—a central focus of marketing. Indeed, many of the metrics throughout this book can be useful in evaluating
brand equity. Certain metrics, however, have been developed specifically to measure
the “health” of brands. This chapter will discuss them.
Although branding strategy is a major aspect of a product offering, there are others,
and managers must be prepared to make trade-offs among them, informed by a sense
of the “worth” of various features. Conjoint analysis helps identify customers’ valuation
of specific product attributes. Increasingly, this technique is used to improve products
and to help marketers evaluate and segment new or rapidly growing markets. In the
final sections of this chapter, we’ll discuss conjoint analysis from multiple perspectives.
Metric
Construction
Considerations
Purpose
4.1
Trial
First-time users
as a percentage
of the target
population.
Distinguish “evertried” from “new”
triers in current
period.
Over time, sales
should rely less on
trial and more on
repeat purchasers.
4.1
Repeat Volume
Repeat buyers,
multiplied by
the number of
products they buy
in each purchase,
multiplied by the
number of times
they purchase per
period.
Depending on
when trial was
achieved, not all
triers will have an
equal opportunity
to make repeat
purchases.
Measure of the
stability of a brand
franchise.
4.1
Penetration
Users in the
previous period,
multiplied by
repeat rate for the
current period,
plus new triers in
the current period.
The length of the
period will affect
norms, that is,
more customers
buy in a year than
in a month.
Measure of the
population buying
in the current
period.
4.1
Volume
Projections
Combine trial
volume and repeat
volume.
Adjust trial and
repeat rates for
time frame. Not all
triers will have time
or opportunity to
repeat.
Plan production
and inventories for
both trade sales
and consumer offtake.
112
MARKETING METRICS
Metric
Construction
Considerations
Purpose
4.2
Year-on-Year
Growth
Percentage change
from one year to
the next.
Distinguish unit
and dollar growth
rates.
Plan production
and budgeting.
4.2
Compound Annual Ending value
May not reflect
Growth Rate
divided by starting individual year-on(CAGR)
value to the power year growth rates.
of 1/N, in which N
is the number of
periods.
4.3
Useful for
averaging growth
rates over long
periods.
Cannibalization
Rate
Percentage of
new product
sales taken from
existing product
line.
Market expansion
effects should also
be considered.
Useful to account
for the fact that
new products
often reduce the
sales of existing
products.
4.3
Fair Share Draw
Assumption that
new entrants
in a market
capture sales
from established
competitors
in proportion
to established
market shares.
May not be
a reasonable
assumption
if there are
significant
differences
among competing
brands.
Useful to generate
an estimate of
sales and shares
after entry of new
competitor.
4.4
Brand Equity
Metrics
Numerous
measures, for
example, Conjoint
utility attributed
to brand.
Metrics tracking
essence of brand
may not track
health and value.
Monitor health of
a brand. Diagnose
weaknesses, as
needed.
4.5
Conjoint Utilities
Regression
coefficients for
attribute levels
derived from
conjoint analysis.
May be function
of number,
level, and type
of attributes in
study.
Indicates the
relative values
that customers
place on attributes
of which product
offerings are
composed.
Continues
Chapter 4 Product and Portfolio Management
113
Metric
Construction
Considerations
Purpose
4.6
Segment Utilities
Clustering of
individuals into
market segments
on the basis of
sum-of-squares
distance between
regression
coefficients drawn
from conjoint
analysis.
May be function of
number, level, and
type of attributes
in conjoint
study. Assumes
homogeneity
within segments.
Uses customer
valuations of
product attributes
to help define
market segments.
4.7
Conjoint Utilities
and Volume
Projection
Used within
Assumes awareness
conjoint simulator and distribution
to estimate volume. levels are known or
can be estimated.
Forecast sales
for alternative
products,
designs, prices, and
branding strategies.
4.1 Trial, Repeat, Penetration, and Volume Projections
Test markets and volume projections enable marketers to forecast sales by sampling
customer intentions through surveys and market studies. By estimating how many
customers will try a new product, and how often they’ll make repeat purchases,
marketers can establish the basis for such projections.
Trial Rate (%) = First-time Triers in Period t (#)
Total Population (#)
First-time Triers in Period t (#) = Total Population (#) * Trial Rate (%)
Penetration t (#) = [Penetration in t-1 (#) * Repeat Rate Period t (%)]
+ First-time Triers in Period t (#)
Projection of Sales t (#) = Penetration t (#) * Average Frequency of Purchase (#)
* Average Units per Purchase (#)
Projections from customer surveys are especially useful in the early stages of
product development and in setting the timing for product launch. Through such
projections, customer response can be estimated without the expense of a full
product launch.
114
MARKETING METRICS
Purpose: To understand volume projections.
When projecting sales for relatively new products, marketers typically use a system
of trial and repeat calculations to anticipate sales in future periods. This works on the
principle that everyone buying the product will either be a new customer (a “trier”) or
a repeat customer. By adding new and repeat customers in any period, we can establish
the penetration of a product in the marketplace.
It is challenging, however, to project sales to a large population on the basis of simulated
test markets, or even full-fledged regional rollouts. Marketers have developed various
solutions to increase the speed and reduce the cost of test marketing, such as stocking a
store with products (or mock-ups of new products) or giving customers money to buy
the products of their choice. These simulate real shopping conditions but require specific models to estimate full-market volume on the basis of test results. To illustrate the
conceptual underpinnings of this process, we offer a general model for making volume
projections on the basis of test market results.
Construction
The penetration of a product in a future period can be estimated on the basis of population size, trial rates, and repeat rates.
Trial Rate (%): The percentage of a defined population that purchases or uses a
product for the first time in a given period.
EXAMPLE: A cable TV company keeps careful records of the names and addresses of its customers. The firm’s vice president of marketing notes that 150 households made first-time use of his
company’s services in March 2009. The company has access to 30,000 households. To calculate the
trial rate for March, we can divide 150 by 30,000, yielding 0.5%.
First-time Triers in Period t (#): The number of customers who purchase or use
a product or brand for the first time in a given period.
Penetration t (#) = [Penetration in t-1 (#) * Repeat Rate Period t (%)]
+ First-time Triers in Period t (#)
EXAMPLE: A cable TV company started selling a monthly sports package in January. The
company typically has an 80% repeat rate and anticipates that this will continue for the new offering. The company sold 10,000 sports packages in January. In February, it expects to add 3,000 customers for the package. On this basis, we can calculate expected penetration for the sports package
in February.
Chapter 4 Product and Portfolio Management
115
Penetration in February = (Penetration January * Repeat Rate)
+ First-time Triers in February
= (10,000 * 80%) + 3,000 = 11,000
Later that year, in September, the company has 20,000 subscribers. Its repeat rate remains 80%. The
company had 18,000 subscribers in August. Management wants to know how many new customers
the firm added for its sports package in September:
First-time Triers = Penetration – Repeat Customers
= 20,000 – (18,000 * 80%) = 5,600
From penetration, it is a short step to projections of sales.
Projection of Sales (#) = Penetration (#) * Frequency of Purchase (#)
* Units per Purchase (#)
Simulated Test Market Results and Volume Projections
Trial Volume
Trial rates are often estimated on the basis of surveys of potential customers. Typically,
these surveys ask respondents whether they will “definitely” or “probably” buy a product. As these are the strongest of several possible responses to questions of purchase
intentions, they are sometimes referred to as the “top two boxes.” The less favorable
responses in a standard five-choice survey include “may or may not buy,” “probably
won’t buy,” and “definitely won’t buy.” (Refer to Section 2.7 for more on intention to
purchase.)
Because not all respondents follow through on their declared purchase intentions,
firms often make adjustments to the percentages in the top two boxes in developing
sales projections. For example, some marketers estimate that 80% of respondents who
say they’ll “definitely buy” and 30% of those who say that they’ll “probably buy” will in
fact purchase a product when given the opportunity.1 (The adjustment for customers
following through is used in the following model.) Although some respondents in the
bottom three boxes might buy a product, their number is assumed to be insignificant.
By reducing the score for the top two boxes, marketers derive a more realistic estimate
of the number of potential customers who will try a product, given the right circumstances. Those circumstances are often shaped by product awareness and availability.
Awareness: Sales projection models include an adjustment for lack of awareness of a
product within the target market (see Figure 4.1). Lack of awareness reduces the trial
rate because it excludes some potential customers who might try the product but don’t
116
MARKETING METRICS
First-Time Use
Repeat Use
Customer Survey
´'HÀQLWHO\%X\µ
´3UREDEO\%X\µ
$GMXVWPHQWIRU&XVWRPHUV
)ROORZLQJ7KURXJK
$GMXVWPHQWIRU$ZDUHQHVV
DQG'LVWULEXWLRQ
7ULDO3RSXODWLRQ
Estimated
Repeat
Rate
Repeat
9ROXPH(DFK
3XUFKDVH
9ROXPH(DFK
3XUFKDVH
Trial Volume
)UHTXHQF\RI
Repeat
3XUFKDVH
+
Repeat Volume
= Total Volume
Figure 4.1
Schematic of Simulated Test Market Volume Projection
know about it. By contrast, if awareness is 100%, then all potential customers know
about the product, and no potential sales are lost due to lack of awareness.
Distribution: Another adjustment to test market trial rates is usually applied—
accounting for the estimated availability of the new product. Even survey respondents
who say they’ll “definitely” try a product are unlikely to do so if they can’t find it easily.
In making this adjustment, companies typically use an estimated distribution, a percentage of total stores that will stock the new product, such as ACV % distribution. (See
Section 6.6 for further detail.)
Adjusted Trial Rate (%) = Trial Rate (%) * Awareness (%) * ACV (%)
Chapter 4 Product and Portfolio Management
117
After making these modifications, marketers can calculate the number of customers
who are expected to try the product, simply by applying the adjusted trial rate to the
target population.
Trial Population (#) = Target Population (#) * Adjusted Trial Rate (%)
Estimated in this way, trial population (#) is identical to penetration (#) in the trial period.
To forecast trial volume, multiply trial population by the projected average number of
units of a product that will be bought in each trial purchase. This is often assumed to be
one unit because most people will experiment with a single unit of a new product before
buying larger quantities.
Trial Volume (#) = Trial Population (#) * Units per Purchase (#)
Combining all these calculations, the entire formula for trial volume is
Trial Volume (#) = Target Population (#) * [(80% * Definitely Buy (#))
+ (30% * Probably Buy (#)) * Awareness (%) * ACV (%)]
* Units per Purchase (#)
EXAMPLE: The marketing team of an office supply manufacturer has a great idea for a new
product—a safety stapler. To sell the idea internally, they want to project the volume of sales they
can expect over the stapler’s first year. Their customer survey yields the following results (see
Table 4.1).
Table 4.1
Customer Survey Responses
% of Customers Responding
Definitely Will Buy
20%
Probably Will Buy
50%
May/May Not Buy
15%
Probably Won’t Buy
10%
Definitely Won’t Buy
5%
Total
100%
On this basis, the company estimates a trial rate for the new stapler by applying the industrystandard expectation that 80% of “definites” and 30% of “probables” will in fact buy the
product if given the opportunity.
118
MARKETING METRICS
Trial Rate = 80% of “Definites” + 30% of “Probables”
= (80% * 20%) + (30% * 50%)
= 31%
Thus, 31% of the population is expected to try the product if they are aware of it and if it is available in stores. The company has a strong advertising presence and a solid distribution network.
Consequently, its marketers believe they can obtain an ACV of approximately 60% for the stapler
and that they can generate awareness at a similar level. On this basis, they project an adjusted trial
rate of 11.16% of the population:
Adjusted Trial Rate = Trial Rate * Awareness * ACV
= 31% * 60% * 60% = 11.16%
The target population comprises 20 million people. The trial population can be calculated by
multiplying this figure by the adjusted trial rate.
Trial Population
= Target Population * Adjusted Trial Rate
= 20 million * 11.16% = 2.232 million
Assuming that each person buys one unit when trying the product, the trial volume will total 2.232
million units.
We can also calculate the trial volume by using the full formula:
Trial Volume = Target Population
* [((80% * Definites) + (30% * Probables)) * Awareness * ACV]
* Units per purchase
= 20m * [((80% * 20%) + (30% * 50%)) * 60% * 60%)] * 1
= 2.232 million
Repeat Volume
The second part of projected volume concerns the fraction of people who try a product
and then repeat their purchase decision. The model for this dynamic uses a single estimated repeat rate to yield the number of customers who are expected to purchase again
after their initial trial. In reality, initial repeat rates are often lower than subsequent
repeat rates. For example, it is not uncommon for 50% of trial purchasers to make a first
repeat purchase, but for 80% of those who purchase a second time to go on to purchase
a third time.
Repeat Buyers (#) = Trial Population (#) * Repeat Rate (%)
To calculate the repeat volume, the repeat buyers figure can then be multiplied by
an expected volume per purchase among repeat customers and by the number of
Chapter 4 Product and Portfolio Management
119
times these customers are expected to repeat their purchases within the period under
consideration.
Repeat Volume (#) = Repeat Buyers (#) * Repeat Unit Volume per Customer (#)
* Repeat Occasions (#)
This calculation yields the total volume that a new product is expected to generate
among repeat customers over a specified introductory period. The full formula can be
written as
Repeat Volume (#) = [Trial Population (#) * Repeat Rate (%)]
* Repeat Unit Volume per Customer (#) * Repeat Occasions (#)
EXAMPLE: Continuing the previous office supplies example, the safety stapler has a trial
population of 2.232 million. Marketers expect the product to be of sufficient quality to generate a
10% repeat rate in its first year. This will yield 223,200 repeat buyers:
Repeat Buyers = Trial Population * Repeat Rate
= 2.232 million * 10%
= 223,200
On average, the company expects each repeat buyer to purchase on four occasions during the first
year. On average, each purchase is expected to comprise two units.
Repeat Volume = Repeat Buyers * Repeat Unit Volume per Customer * Repeat Occasions
= 223,200 * 2 * 4
= 1,785,600 units
This can be represented in the full formula:
Repeat Volume (#) = [Repeat Rate (%) * Trial Population (#)]
* Repeat Volume per Customer (#)
* Repeat Occasions (#)
= (10% * 2,232,000) * 2 * 4
= 1,785,600 units
Total Volume
Total volume is the sum of trial volume and repeat volume, as all volume must be sold
to either new customers or returning customers.
Total Volume (#) = Trial Volume (#) + Repeat Volume (#)
120
MARKETING METRICS
To capture total volume in its fully detailed form, we need only combine the previous
formulas.
Total Volume (#) = [Target Population * ((0.8 * Definitely Buy + 0.3 * Probably Buy)
* Awareness * ACV) * Units per Trial Purchase]
+ [(Trial Population * Repeat Rate)
* Repeat Volume per Customer * Repeat Occasions]
EXAMPLE:
volume.
Total volume in year one for the stapler is the sum of trial volume and repeat
Total Volume = Trial Volume + Repeat Volume
= 2,232,000 + 1,785,600
= 4,017,600 Units
A full calculation of this figure and a template for a spreadsheet calculation are presented in
Table 4.2.
Table 4.2
Volume Projection Spreadsheet
Preliminary Data
Source
Definitely Will Buy
Customer Survey
20%
Probably Will Buy
Customer Survey
50%
Likely Buyers from Definites
= Definitely Buy * 80%
16%
Likely Buyers from Probables
= Probably Buy * 30%
15%
Trial Rate (%)
Total of Likely Buyers
31%
Awareness
Estimated from Marketing Plan
60%
ACV
Estimated from Marketing Plan
60%
Adjusted Trial Rate (%)
= Trial Rate * Awareness * ACV
11.2%
Target Population (#) (thousands)
Marketing Plan Data
20,000
Trial Population (#) (thousands)
= Target Population * Adjusted
Trial Rate
2,232
Likely Buyers
Marketing Adjustments
Continues
Chapter 4 Product and Portfolio Management
121
Table 4.2
Continued
Preliminary Data
Source
Unit Volume Purchased per Trial (#)
Estimated from Marketing Plan
1
Trial Volume (#) (Thousands)
= Trial Population * Volume per
Trier
2,232
Repeat Rate (%)
Estimated from Marketing Plan
10%
Repeat Buyers (#)
= Repeat Rate * Trial Population
223,200
Avg. Volume per Repeat
Purchase (#)
Estimated from Marketing Plan
2
Repeat Purchase Frequency ** (#)
Estimated from Marketing Plan
4
Repeat Volume (#) (Thousands)
= Repeat Buyers * Repeat Volume
per Purchase * Repeat Purchase
1,786
Total Volume (#) (Thousands)
4,018
**Note: The average frequency of repeat purchases per repeat purchaser should be adjusted to reflect the
time available for first-time triers to repeat, the purchase cycle (frequency) for the category, and availability. For example, if trial rates are constant over the year, the number of repeat purchases would be about
50% of what it would have been if all had tried on day 1 of the period.
Data Sources, Complications, and Cautions
Sales projections based on test markets will always require the inclusion of key assumptions. In setting these assumptions, marketers face tempting opportunities to make the
assumptions fit the desired outcome. Marketers must guard against that temptation
and perform sensitivity analysis to establish a range of predictions.
Relatively simple metrics such as trial and repeat rates can be difficult to capture in
practice. Although strides have been made in gaining customer data—through customer loyalty cards, for example—it will often be difficult to determine whether
customers are new or repeat buyers.
Regarding awareness and distribution: Assumptions concerning the level of public
awareness to be generated by launch advertising are fraught with uncertainty. Marketers
are advised to ask: What sort of awareness does the product need? What complementary
promotions can aid the launch?
122
MARKETING METRICS
Trial and repeat rates are both important. Some products generate strong results in the
trial stage but fail to maintain ongoing sales. Consider the following example.
EXAMPLE: Let’s compare the safety stapler with a new product, such as an enhanced envelope
sealer. The envelope sealer generates less marketing buzz than the stapler but enjoys a greater repeat
rate. To predict results for the envelope sealer, we have adapted the data from the safety stapler by
reducing the top two box responses by half (reflecting its lower initial enthusiasm) and raising the
repeat rate from 10% to 33% (showing stronger product response after use).
At the six-month mark, sales results for the safety stapler are superior to those for the envelope
sealer. After one year, sales results for the two products are equal. On a three-year time scale, however, the envelope sealer—with its loyal base of customers—emerges as the clear winner in sales
volume (see Figure 4.2).
The data for the graph is derived as shown in Table 4.3.
Volume of Sales Generated (Thousands)
Volume Over Time: High Interest Versus Loyalty Generating Products
7,000
6,000
5,000
4,000
3,000
2,000
6 Months
12 Months
18 Months
2 Years
Safety Stapler
3,125
4,018
4,910
5,803
Envelope Sealer
2,589
4,062
5,535
7,008
Time from Product Launch
Figure 4.2 Time Horizon Influences Perceived Results
Chapter 4 Product and Portfolio Management
123
124
Table 4.3
High Initial Interest or Long-Term Loyalty—Results over Time
6 Months
12 Months
2 Years
18 Months
MARKETING METRICS
Preliminary Data
Source
Prod A
Prod B Prod A
Prod B Prod A
Prod BProd A
Definitely Will Buy
Customer Survey
20%
10%
20%
10%
20%
10%
20%
10%
Probably Will Buy
Customer Survey
50%
25%
50%
25%
50%
25%
50%
25%
= Definitely Buy
* 80%
16%
8%
16%
8%
16%
8%
16%
8%
Likely Buyers from Probables = Probably Buy
* 30%
15%
8%
15%
8%
15%
8%
15%
8%
Trial Rate
Total of Likely
Buyers
31%
16%
31%
16%
31%
16%
31%
16%
Awareness
Estimated from
Marketing Plan
60%
60%
60%
60%
60%
60%
60%
60%
ACV
Estimated from
Marketing Plan
60%
60%
60%
60%
60%
60%
60%
60%
Adjusted Trial Rate
= Trial Rate *
Awareness * ACV
11.2%
5.6%
11.2%
5.6%
11.2%
5.6%
11.2%
5.6%
Prod B
Differences Highlighted in
Yellow
Likely Buyers
Likely Buyers from Definites
Marketing Adjustments
Target Population (Thousands)Marketing Plan Data 20,000
20,000 20,000
20,000 20,000
20,000 20,000
20,000
Chapter 4 Product and Portfolio Management
Trial Population (Thousands) = Target Population * 2,232
Adjusted Trial Rate
1,116
2,232
1,116
2,232
1,116 2,232
1,116
Unit Volume Purchased
at Trial
Estimated from
Marketing Plan
1
1
1
1
1
1
1
Trial Volume (Thousands)
= Trial Population *
Volume bought
2,232
1,116
2,232
1,116
2,232
1,116 2,232
1,116
Repeat Rate
Estimated from
Marketing Plan
10%
33%
10%
33%
10%
33%
33%
Repeat Buyers
= Repeat Rate * Trial 223.20
Population
1
10%
368.28 223.20
368.28 223.20
368.28 223.20
368.28
Repeat Purchase Unit Volume Estimated from
Marketing Plan
2
2
2
2
2
2
2
2
Number of Repeat Purchases
Estimated from
Marketing Plan
2
2
4
4
6
6
8
8
Repeat Volume (Thousands)
= Repeat Buyers *
Repeat Volume *
Number of Repeat
Purchases
893
1,473
1,786
2,946
2,678
4,419 3,571
5,892
5,535
7,008
Total Volume
3,125
2,589
4,018
4,062
4,910
5,803
125
Repeating and Trying: Some models assume that customers, after they stop repeating
purchases, are lost and do not return. However, customers may be acquired, lost, reacquired, and lost again. In general, the trial-repeat model is best suited to projecting sales
over the first few periods. Other means of predicting volume include share of requirements and penetration metrics (refer to Sections 2.4 and 2.5). Those approaches may be
preferable for products that lack reliable repeat rates.
Market
Size
Penetration Share of
Usage
Share
Requirements Index
Market
Share
Units
Sold
New
Product
1,000,000
5%
80%
1.2
4.8%
48,000
Source
Estimated
Estimated
Estimated
Estimated
Penetration
Share *
Share *
Market
Share of
Size
Requirements
* Usage Index
Related Metrics and Concepts
Ever-Tried: This is slightly different from trial in that it measures the percentage of the
target population that has “ever” (in any previous period) purchased or consumed the
product under study. Ever-tried is a cumulative measure and can never add up to more
than 100%. Trial, by contrast, is an incremental measure. It indicates the percentage
of the population that tries the product for the first time in a given period. Even here,
however, there is potential for confusion. If a customer stops buying a product but tries
it again six months later, some marketers will categorize that individual as a returning
purchaser, others as a new customer. By the latter definition, if individuals can “try” a
product more than once, then the sum of all “triers” could equal more than the total
population. To avoid confusion, when reviewing a set of data, it’s best to clarify the
definitions behind it.
Variations on Trial: Certain scenarios reduce the barriers to trial but entail a lower
commitment by the customer than a standard purchase.
■
■
126
Forced Trial: No other similar product is available. For example, many people
who prefer Pepsi-Cola have “tried” Coca-Cola in restaurants that only serve the
latter, and vice versa.
Discounted Trial: Consumers buy a new product but at a substantially reduced
price.
MARKETING METRICS
Forced and discounted trials are usually associated with lower repeat rates than trials
made through volitional purchase.
Evoked Set: The set of brands that consumers name in response to questions about
which brands they consider (or might consider) when making a purchase in a specific
category. Evoked Sets for breakfast cereals, for example, are often quite large, while
those for coffee may be smaller.
Number of New Products: The number of products introduced for the first time in a
specific time period.
Revenue from New Products: Usually expressed as the percentage of sales generated
by products introduced in the current period or, at times, in the most recent three to
five periods.
Margin on New Products: The dollar or percentage profit margin on new products.
This can be measured separately but does not differ mathematically from margin
calculations.
Company Profit from New Products: The percentage of company profits that is
derived from new products. In working with this figure, it is important to understand
how “new product” is defined.
Target Market Fit: Of customers purchasing a product, target market fit represents the
percentage who belong in the demographic, psychographic, or other descriptor set for that
item. Target market fit is useful in evaluating marketing strategies. If a large percentage of
customers for a product belongs to groups that have not previously been targeted, marketers may reconsider their targets—and their allocation of marketing spending.
4.2 Growth: Percentage and CAGR
There are two common measures of growth. Year-on-year percentage growth uses
the prior year as a base for expressing percentage change from one year to the next.
Over longer periods of time, compound annual growth rate (CAGR) is a generally
accepted metric for average growth rates.
Year-on-Year Growth (%) =
Value ($,#,%) t – Value ($,#,%) t – 1
Value ($,#,%) t – 1
Compound Annual Growth Rate, or = {[Ending Value ($,#,%)/Starting Value ($,#,%)]
CAGR (%)
^ [1/Number of Years (#)]} – 1
Same stores growth = Growth calculated only on the basis of stores that were
fully established in both the prior and current periods.
Chapter 4 Product and Portfolio Management
127
Purpose: To measure growth.
Growth is the aim of virtually all businesses. Indeed, perceptions of the success or failure of many enterprises are based on assessments of their growth. Measures of year-onyear growth, however, are complicated by two factors:
1. Changes over time in the base from which growth is measured. Such changes
might include increases in the number of stores, markets, or salespeople generating sales. This issue is addressed by using “same store” measures (or corollary
measures for markets, sales personnel, and so on).
2. Compounding of growth over multiple periods. For example, if a company
achieves 30% growth in one year, but its results remain unchanged over the
two subsequent years, this would not be the same as 10% growth in each of
three years. CAGR, the compound annual growth rate, is a metric that addresses this issue.
Construction
Percentage growth is the central plank of year-on-year analysis. It addresses the question: What has the company achieved this year, compared to last year? Dividing the
results for the current period by the results for the prior period will yield a comparative
figure. Subtracting one from the other will highlight the increase or decrease between
periods. When evaluating comparatives, one might say that results in Year 2 were, for
example, 110% of those in Year 1. To convert this figure to a growth rate, one need only
subtract 100%.
The periods considered are often years, but any time frame can be chosen.
Year-on-Year Growth (%) =
Value ($,#,%) t – Value ($,#,%) t – 1
Value ($,#,%) t – 1
EXAMPLE: Ed’s is a small deli, which has had great success in its second year of operation.
Revenues in Year 2 are $570,000, compared with $380,000 in Year 1. Ed calculates his second-year
sales results to be 150% of first-year revenues, indicating a growth rate of 50%.
Year-on-Year Sales Growth =
$570,000 – $380,000
$380,000
= 50%
Same Stores Growth: This metric is at the heart of retail analysis. It enables marketers to analyze results from stores that have been in operation for the entire period under
128
MARKETING METRICS
consideration. The logic is to eliminate the stores that have not been open for the full
period to ensure comparability. Thus, same stores growth sheds light on the effectiveness with which equivalent resources were used in the period under study versus the
prior period. In retail, modest same stores growth and high general growth rates would
indicate a rapidly expanding organization, in which growth is driven by investment.
When both same stores growth and general growth are strong, a company can be
viewed as effectively using its existing base of stores.
EXAMPLE: A small retail chain in Bavaria posts impressive percentage growth figures, moving from €58 million to €107 million in sales (84% growth) from one year to the next. Despite this
dynamic growth, however, analysts cast doubt on the firm’s business model, warning that its same
stores growth measure suggests that its concept is failing (see Table 4.4).
Table 4.4 Revenue of a Bavarian Chain Store
Store
Opened
Revenue First Year (m)
Revenue Second Year (m)
A
Year 1
€10
€9
B
Year 1
€19
€20
C
Year 1
€20
€15
D
Year 1
€9
€11
E
Year 2
n/a
€52
€58
€107
Same stores growth excludes stores that were not open at the beginning of the first year under
consideration. For simplicity, we assume that stores in this example were opened on the first
day of Years 1 and 2, as appropriate. On this basis, same stores revenue in Year 2 would be €55
million—that is, the €107 million total for the year, less the €52 million generated by the newly
opened Store E. This adjusted figure can be entered into the same stores growth formula:
Same Stores Growth =
=
(Stores A-D Sales Year 2) – (Stores A-D Sales Year 1)
$Stores A-D Sales Year 1
€55m – €58m
€58
= –5%
Chapter 4 Product and Portfolio Management
129
As demonstrated by its negative same stores growth figure, sales growth at this firm has been fueled
entirely by a major investment in a new store. This suggests serious doubts about its existing store
concept. It also raises a question: Did the new store “cannibalize” existing store sales? (See the next
section for cannibalization metrics.)
Compounding Growth, Value at Future Period: By compounding, managers adjust
growth figures to account for the iterative effect of improvement. For example, 10%
growth in each of two successive years would not be the same as a total of 20% growth
over the two-year period. The reason: Growth in the second year is built upon the
elevated base achieved in the first. Thus, if sales run $100,000 in Year 0 and rise by 10%
in Year 1, then Year 1 sales come to $110,000. If sales rise by a further 10% in Year 2,
however, then Year 2 sales do not total $120,000. Rather, they total $110,000 + (10% *
$110,000) = $121,000.
The compounding effect can be easily modeled in spreadsheet packages, which enable
you to work through the compounding calculations one year at a time. To calculate a
value in Year 1, multiply the corresponding Year 0 value by one plus the growth rate.
Then use the value in Year 1 as a new base and multiply it by one plus the growth rate to
determine the corresponding value for Year 2. Repeat this process through the required
number of years.
EXAMPLE:
$133.10.
Over a three-year period, $100 dollars, compounded at a 10% growth rate, yields
Year 0 to Year 1 $100 + 10% Growth (that is, $10) = $110
Year 1 to Year 2 $110 + 10% Growth ($11) = $121
Year 2 to Year 3 $121 + 10% Growth ($12.10) = $133.10
There is a mathematical formula that generates this effect. It multiplies the value at the
beginning—that is, in Year 0—by one plus the growth rate to the power of the number
of years over which that growth rate applies.
Value in Future Period ($,#,%) = Current Value ($,#,%) * [(1 + CAGR (%)) ^ Number of
Periods (#)]
EXAMPLE: Using the formula, we can calculate the impact of 10% annual growth over a
period of three years. The value in Year 0 is $100. The number of years is 3. The growth rate is 10%.
130
MARKETING METRICS
Value in Future Period =Value in Year 0 * (1 + Growth Rate) ^ Number of Years
=$100 * (100% + 10%) ^ 3
=$100 * 133.1% = $133.10
Compound Annual Growth Rate (CAGR): The CAGR is a constant year-on-year
growth rate applied over a period of time. Given starting and ending values, and the
length of the period involved, it can be calculated as follows:
CAGR (%) = {[Ending Value ($,#)/Starting Value ($,#)] ^ 1/Number of Periods (#)} – 1
EXAMPLE: Let’s assume we have the results of the compounding growth observed in the
previous example, but we don’t know what the growth rate was. We know that the starting value
was $100, the ending value was $133.10, and the number of years was 3. We can simply enter these
numbers into the CAGR formula to derive the CAGR.
CAGR = [(Ending Value/Starting Value) ^ (1/Number of Years)] – 1
= [($133.10/$100) ^ 1/3] – 1
= [1.331 (The Increase) ^ 1/3 (Cube Root)] – 1 = 1.1 – 1 = 10%
Thus, we determine that the growth rate was 10%.
Data Sources, Complications, and Cautions
Percentage growth is a useful measure as part of a package of metrics. It can be deceiving, however, if not adjusted for the addition of such factors as stores, salespeople, or
products, or for expansion into new markets. “Same store” sales, and similar adjustments for other factors, tell us how effectively a company uses comparable resources.
These adjustments, however, are limited by their deliberate omission of factors that
weren’t in operation for the full period under study. Adjusted figures must be reviewed
in tandem with measures of total growth.
Related Metrics and Concepts
Life Cycle: Marketers view products as passing through four stages of development:
■
Introductory: Small markets not yet growing fast.
■
Growth: Larger markets with faster growth rates.
■
Mature: Largest markets but little or no growth.
■
Decline: Variable size markets with negative growth rates.
This is a rough classification. No generally accepted rules exist for making these
classifications.
Chapter 4 Product and Portfolio Management
131
4.3 Cannibalization Rates and Fair Share Draw
Cannibalization is the reduction in sales (units or dollars) of a firm’s existing products due to the introduction of a new product. The cannibalization rate is generally
calculated as the percentage of a new product’s sales that represents a loss of sales
(attributable to the introduction of the new entrant) of a specific existing product
or products.
Cannibalization Rate (%) =
Sales Lost from Existing Products (#,$)
Sales of New Product (#,$)
Cannibalization rates represent an important factor in the assessment of new
product strategies.
Fair share draw constitutes an assumption or expectation that a new product will
capture sales (in unit or dollar terms) from existing products in proportion to
the market shares of those existing products.
Cannibalization is a familiar business dynamic. A company with a successful product
that has strong market share is faced by two conflicting ideas. The first is that it wants
to maximize profits on its existing product line, concentrating on the current strengths
that promise success in the short term. The second idea is that this company—or its
competitors—may identify opportunities for new products that better fit the needs of
certain segments. If the company introduces a new product in this field, however, it
may “cannibalize” the sales of its existing products. That is, it may weaken the sales of
its proven, already successful product line. If the company declines to introduce the
new product, however, it will leave itself vulnerable to competitors who will launch
such a product, and may thereby capture sales and market share from the company.
Often, when new segments are emerging and there are advantages to being early to
market, the key factor becomes timing. If a company launches its new product too
early, it may lose too much income on its existing line; if it launches too late, it may miss
the new opportunity altogether.
Cannibalization: A market phenomenon in which sales of one product are
achieved at the expense of some of a firm’s other products.
The cannibalization rate is the percentage of sales of a new product that come from a
specific set of existing products.
Cannibalization Rate (%) =
Sales Lost from Existing Products (#,$)
Sales of New Product (#,$)
132
MARKETING METRICS
EXAMPLE: A company has a single product that sold 10 units in the previous period. The
company plans to introduce a new product that will sell 5 units with a cannibalization rate of 40%.
Thus 40% of the sales of the new product (40% * 5 units = 2 units) come at the expense of the old
product. Therefore, after cannibalization, the company can expect to sell 8 units of the old product
and 5 of the new product, or 13 units in total.
Any company considering the introduction of a new product should confront the
potential for cannibalization. A firm would do well to ensure that the amount of cannibalization is estimated beforehand to provide an idea of how the product line’s contribution as a whole will change. If performed properly, this analysis will tell a company
whether overall profits can be expected to increase or decrease with the introduction of
the new product line.
EXAMPLE: Lois sells umbrellas on a small beach where she is the only provider. Her financials
for last month were as follows:
Umbrella Sales Price:
$20
Variable Cost per Umbrella:
$10
Umbrella Contribution per Unit:
$10
Total Unit Sales per Month:
100
Total Monthly Contribution:
$1,000
Next month, Lois plans to introduce a bigger, lighter-weight umbrella called the “Big Block.”
Projected financials for the Big Block are as follows:
Big Block Sales Price:
$30
Variable Cost per Big Block:
$15
Big Block Contribution per Unit:
$15
Total Unit Sales per Month (Big Block):
50
Total Monthly Contribution (Big Block):
$750
If there is no cannibalization, Lois thus expects her total monthly contribution will be $1,000 + $750,
or $1,750. Upon reflection, however, Lois thinks that the unit cannibalization rate for Big Block will
be 60%. Her projected financials after accounting for cannibalization are therefore as follows:
Big Block Unit Sales:
50
Cannibalization Rate:
60%
Regular Umbrella Sales Lost:
50 * 60% = 30
Chapter 4 Product and Portfolio Management
133
New Regular Umbrella Sales:
100 – 30 = 70
New Total Contribution (Regular):
70 Units * $10 Contribution per Unit = $700
Big Block Total Contribution:
50 Units * $15 Contribution per Unit = $750
Lois’ Total Monthly Contribution:
$1,450
Under these projections, total umbrella sales will increase from 100 to 120, and total contribution
will increase from $1,000 to $1,450. Lois will replace 30 regular sales with 30 Big Block sales and
gain an extra $5 unit contribution on each. She will also sell 20 more umbrellas than she sold last
month and gain $15 unit contribution on each.
In this scenario, Lois was in the enviable position of being able to cannibalize a lower-margin product with a higher-margin one. Sometimes, however, new products carry unit contributions lower
than those of existing products. In these instances, cannibalization reduces overall profits for the
firm.
An alternative way to account for cannibalization is to use a weighted contribution
margin. In the previous example, the weighted contribution margin would be the unit
margin Lois receives for Big Block after accounting for cannibalization. Because each
Big Block contributes $15 directly and cannibalizes the $10 contribution generated
by regular umbrellas at a 60% rate, Big Block’s weighted contribution margin is $15
– (0.6 * $10), or $9 per unit. Because Lois expects to sell 50 Big Blocks, her total contribution is projected to increase by 50 * $9, or $450. This is consistent with our previous
calculations.
If the introduction of Big Block requires some fixed marketing expenditure, then the
$9 weighted margin can be used to find the break-even number of Big Block sales
required to justify that expenditure. For example, if the launch of Big Block requires
$360 in one-time marketing costs, then Lois needs to sell $360/$9, or 40 Big Blocks to
break even on that expenditure.
If a new product has a margin lower than that of the existing product that it cannibalizes, and if its cannibalization rate is high enough, then its weighted contribution margin might be negative. In that case, company earnings will decrease with each unit of
the new product sold.
Cannibalization refers to a dynamic in which one product of a firm takes share from
one or more other products of the same firm. When a product takes sales from a
competitor’s product, that is not cannibalization . . . though managers sometimes incorrectly state that their new products are “cannibalizing” sales of a competitor’s goods.
Though it is not cannibalization, the impact of a new product on the sales of competing goods is an important consideration in a product launch. One simple assumption
about how the introduction of a new product might affect the sales of existing products
is called “fair share draw.”
134
MARKETING METRICS
Fair Share Draw: The assumption that a new product will capture sales (in unit
or dollar terms) from existing products in direct proportion to the market shares
held by those existing products.
EXAMPLE: Three rivals compete in the youth fashion market in a small town. Their sales and
market shares for last year appear in the following table.
Firm
Sales
Share
Threadbare
$500,000
50%
Too Cool for School
$300,000
30%
Tommy Hitchhiker
$200,000
20%
Total
$1,000,000
100%
A new entrant is expected to enter the market in the coming year and to generate $300,000 in sales.
Two-thirds of those sales are expected to come at the expense of the three established competitors.
Under an assumption of fair share draw, how much will each firm sell next year?
If the new firm takes two-thirds of its sales from existing competitors, then this “capture” of sales
will total (2/3) * $300,000, or $200,000. Under fair share draw, the breakdown of that $200,000 will
be proportional to the shares of the current competitors. Thus 50% of the $200,000 will come from
Threadbare, 30% from Too Cool, and 20% from Tommy. The following table shows the projected
sales and market shares next year of the four competitors under the fair share draw assumption:
Firm
Sales
Share
Threadbare
$400,000
36.36%
Too Cool for School
$240,000
21.82%
Tommy Hitchhiker
$160,000
14.55%
New Entrant
$300,000
27.27%
Total
$1,100,000
100%
Notice that the new entrant expands the market by $100,000, an amount equal to the sales of
the new entrant that do not come at the expense of existing competitors. Notice also that under
fair share draw, the relative shares of the existing competitors remain unchanged. For example,
Threadbare’s share, relative to the total of the original three competitors, is 36.36/(36.36 + 21.82 +
14.55), or 50%—equal to its share before the entry of the new competitor.
Chapter 4 Product and Portfolio Management
135
The opposite of cannibalization is incremental sales. This is when the introduction of
a new product may boost sales for a complementary product—one that naturally goes
with the product.
Data Sources, Complications, and Cautions
As noted previously, in cannibalization, one of a firm’s products takes sales from one
or more of that firm’s other products. Sales taken from the products of competitors are
not “cannibalized” sales, though some managers label them as such.
Cannibalization rates depend on how the features, pricing, promotion, and distribution of the new product compare to those of a firm’s existing products. The greater the
similarity of their respective marketing strategies, the higher the cannibalization rate is
likely to be.
Although cannibalization is always an issue when a firm launches a new product that
competes with its established line, this dynamic is particularly damaging to the firm’s
profitability when a low-margin entrant captures sales from the firm’s higher-margin
offerings. In such cases, the new product’s weighted contribution margin can be negative. Even when cannibalization rates are significant, however, and even if the net effect
on the bottom line is negative, it may be wise for a firm to proceed with a new product
if management believes that the original line is losing its competitive strength. The
following example is illustrative.
EXAMPLE: A producer of powdered-milk formula has an opportunity to introduce a new,
improved formula. The new formula has certain attributes not found in the firm’s existing products. Due to higher costs, however, it will carry a contribution margin of only $8, compared with
the $10 margin of the established formula. Analysis suggests that the unit cannibalization rate of the
new formula will be 90% in its initial year. If the firm expects to sell 300 units of the new formula in
its first year, should it proceed with the introduction?
Analysis shows that the new formula will generate $8 * 300, or $2,400 in direct contribution.
Cannibalization, however, will reduce contribution from the established line by $10 * 0.9 * 300,
or $2,700. Thus, the company’s overall contribution will decline by $300 with the introduction of
the new formula. (Note also that the weighted unit margin for the new product is –$1.) This simple
analysis suggests that the new formula should not be introduced.
The following table, however, contains the results of a more detailed four-year analysis. Reflected in
this table are management’s beliefs that without the new formula, sales of the regular formula will
decline to 700 units in Year 4. In addition, unit sales of the new formula are expected to increase to
600 in Year 4, while cannibalization rates decline to 60%.
136
MARKETING METRICS
Unit Sales of Regular Formula
Without New Product Launch
Year 1
Year 2
Year 3
Year 4
Total
1,000
900
800
700
3,400
—
—
Unit Sales of New Formula
300
400
500
600
1,800
Cannibalization Rate
90%
80%
70%
60%
—
Unit Sales of Regular Formula
with New Product Launch
730
580
450
340
2,100
Without the new formula, total four-year contribution is projected as $10 * 3,400, or $34,000.
With the new formula, total contribution is projected as ($8 * 1,800) + ($10 * 2,100), or $35,400.
Although forecast contribution is lower in Year 1 with the new formula than without it, total fouryear contribution is projected to be higher with the new product due to increases in new-formula
sales and decreases in the cannibalization rate.
4.4 Brand Equity Metrics
Brand equity is strategically crucial, but famously difficult to quantify. Many experts
have developed tools to analyze this asset, but there’s no universally accepted way to
measure it. In this section, we’ll consider the following techniques to gain insight in
this area:
Brand Equity Ten (Aaker)
Brand Asset® Valuator (Young & Rubicam)
Brand Equity Index (Moran)
Brand Finance
BrandZ (Millward Brown)
Brand Valuation Model (Interbrand)
Purpose: To measure the value of a brand.
A brand encompasses the name, logo, image, and perceptions that identify a product,
service, or provider in the minds of customers. It takes shape in advertising, packaging,
and other marketing communications, and becomes a focus of the relationship with
Chapter 4 Product and Portfolio Management
137
consumers. In time, a brand comes to embody a promise about the goods it identifies—
a promise about quality, performance, or other dimensions of value, which can influence consumers’ choices among competing products. When consumers trust a brand
and find it relevant, they may select the offerings associated with that brand over those
of competitors, even at a premium price. When a brand’s promise extends beyond a
particular product, its owner may leverage it to enter new markets. For all these reasons,
a brand can hold tremendous value, known as brand equity.
Yet this value can be remarkably difficult to measure. At a corporate level, when one
company buys another, marketers might analyze the goodwill component of the purchase price to shed light on the value of the brands acquired. As goodwill represents
the excess paid for a firm—beyond the value of its tangible, measurable assets, and as
a company’s brands constitute important intangible assets—the goodwill figure may
provide a useful indicator of the value of a portfolio of brands. Of course, a company’s
brands are rarely the only intangible assets acquired in such a transaction. Goodwill
more frequently encompasses intellectual property and other intangibles in addition to
brand. The value of intangibles, as estimated by firm valuations (sales or share prices),
is also subject to economic cycles, investor “exuberance,” and other influences that are
difficult to separate from the intrinsic value of the brand.
From a consumer’s perspective, the value of a brand might be the amount she would be
willing to pay for merchandise that carries the brand’s name, over and above the price
she’d pay for identical unbranded goods.2 Marketers strive to estimate this premium in
order to gain insight into brand equity. Here again, however, they encounter daunting
complexities, as individuals vary not only in their awareness of different brands, but
in the criteria by which they judge them, the evaluations they make, and the degree to
which those opinions guide their purchase behavior. In a similar vein, the value of the
brand might be the increased propensity to buy the product because it carries the brand
name.
Theoretically, a marketer might aggregate these preferences across an entire population
to estimate the total premium its members would pay for goods of a certain brand. Even
that, however, wouldn’t fully capture brand equity. What’s more, the value of a brand
encompasses not only the premium a customer will pay for each unit of merchandise
associated with that brand, but also the incremental volume it generates. A successful
brand will shift outward the demand curve for its goods or services; that is, it not only
will enable a provider to charge a higher price (P' rather than P, as seen in Figure 4.3),
but it will also sell an increased quantity (Q' rather than Q). Thus, brand equity in this
example can be viewed as the difference between the revenue with the brand (P' x Q')
and the revenue without the brand (P x Q)—depicted as the shaded area in Figure 4.3.
(Of course, this example focuses on revenue, when, in fact, it is profit or present value
of profits that matters more. But, given the increase in price associated with the brand
comes with no incremental variable costs, and, as such, this lift represents profits,
as well.)
138
MARKETING METRICS
Price
P'
P
High Brand Equity
Low Brand Equity
Q
Figure 4.3
Q'
Quantity
Brand Equity—Outward Shift of Demand Curve
In practice, of course, it’s difficult to measure a demand curve, and few marketers do
so. Because brands are crucial assets, however, both marketers and academic researchers have devised means to contemplate their value. David Aaker, for example, tracks 10
attributes of a brand to assess its strength. Bill Moran has formulated a brand equity
index that can be calculated as the product of effective market share, relative price, and
customer retention. Kusum Ailawadi and her colleagues have refined this calculation,
suggesting that a truer estimate of a brand’s value might be derived by multiplying
the Moran index by the dollar volume of the market in which it competes. Young &
Rubicam, a marketing communications agency, has developed a tool called the Brand
Asset Valuator©, which measures a brand’s power on the basis of differentiation, relevance, esteem, and knowledge. An even more theoretical conceptualization of brand
equity is the difference of the firm value with and without the brand. If you find it difficult to imagine the firm without its brand, then you can appreciate how difficult it
is to quantify brand equity. Interbrand, a brand strategy agency, draws upon its own
model to separate tangible product value from intangible brand value and uses the latter to rank the top 100 global brands each year. Finally, conjoint analysis can shed light
on a brand’s value because it enables marketers to measure the impact of that brand on
customer preference, treating it as one among many attributes that consumers trade off
in making purchase decisions (see Section 4.5).
Construction
Brand Equity Ten (Aaker): David Aaker, a marketing professor and brand consultant,
highlights 10 attributes of a brand that can be used to assess its strength. These include
Differentiation, Satisfaction or Loyalty, Perceived Quality, Leadership or Popularity,
Perceived Value, Brand Personality, Organizational Associations, Brand Awareness,
Market Share, and Market Price and Distribution Coverage. Aaker doesn’t weight the
attributes or combine them in an overall score, as he believes any weighting would be
Chapter 4 Product and Portfolio Management
139
arbitrary and would vary among brands and categories. Rather, he recommends tracking each attribute separately.
Brand Equity Index (Moran): Marketing executive Bill Moran has derived an index of
brand equity as the product of three factors: Effective Market Share, Relative Price, and
Durability.
Brand Equity Index (I) = Effective Market Share (%) * Relative Price (I) * Durability (%)
Effective Market Share is a weighted average. (We suggest that it makes more sense
to use market share in units rather than in revenue terms, to avoid effectively double
counting the impact of price, which also influences the calculation through “Relative
Price.”) It represents the sum of a brand’s market shares in all segments in which it
competes, weighted by each segment’s proportion of that brand’s total sales. Thus, if a
brand made 70% of its sales in Segment A, in which it had a 50% share of the market,
and 30% of its sales in Segment B, in which it had a 20% share, its Effective Market
Share would be (0.7 * 0.5) + (0.3 * 0.2) = 0.35 + 0.06 = 0.41, or 41%.
Relative Price is a ratio. It represents the price of goods sold under a given brand,
divided by the average price of comparable goods in the market. For example, if goods
associated with the brand under study sold for $2.50 per unit, while competing goods
sold for an average of $2.00, that brand’s Relative Price would be 1.25, and it would be
said to command a price premium. Conversely, if the brand’s goods sold for $1.50, versus $2.00 for the competition, its Relative Price would be 0.75, placing it at a discount to
the market. Note that this measure of relative price is not the same as dividing the brand
price by the market average price. It does have the advantage that, unlike the latter, the
calculated value is not affected by the market share of the firm or its competitors.
Durability is a measure of customer retention or loyalty. It represents the percentage of
a brand’s customers who will continue to buy goods under that brand in the following
year.
EXAMPLE: ILLI is a tonic drink that focuses on two geographic markets—eastern and western
U.S. metropolitan areas. In the western market, which accounts for 60% of ILLI’s sales, the drink
has a 30% share of the market. In the East, where ILLI makes the remaining 40% of its sales, it has a
50% share of the market.
Effective Market Share is equal to the sum of ILLI’s shares of the segments, weighted by the percentage of total brand sales represented by each.
West = 30% * 60% = 0.18
East = 50% * 40% = 0.20
Effective Market Share = 0.38
140
MARKETING METRICS
The average price for tonic drinks is $2.00, but ILLI enjoys a premium. It generally sells for $2.50,
yielding a Relative Price of $2.50 / $2.00, or 1.25.
Half of the people who purchase ILLI this year are expected to repeat next year, generating a
Durability figure of 0.5. (See Section 4.1 for a definition of repeat rates.)
With this information, ILLI’s Brand Equity Index can be calculated as follows:
Brand Equity = Effective Market Share * Relative Price * Durability = 0.38*1.25*0.5
= 0.2375
Clearly, marketers can expect to encounter interactions among the three factors behind
a Brand Equity Index. If they raise the price of a brand’s goods, for example, they may
increase its Relative Price but reduce its Effective Market Share and Durability. Would
the overall effect be positive for the brand? By estimating the Brand Equity Index before
and after the price increase under consideration, marketers may gain insight into that
question.
Notice that two of the factors behind this index, Effective Market Share and Relative
Price, draw upon the axes of a demand curve (quantity and price). In constructing his
index, Moran has taken those two factors and combined them, through year-to-year
retention, with the dimension of time.
Ailawadi and her colleagues suggested that the equity index of a brand can be enhanced
by multiplying it by the dollar volume of the market in which the brand competes, generating a better estimate of its value. Ailawadi also contends that the equity of a brand is
better captured by its overall revenue premium (relative to generic goods) rather than
its price per unit alone, as the revenue figure incorporates both price and quantity and
so reflects a jump from one demand curve to another rather than a movement along a
single curve.
Brand Asset Valuator (Young & Rubicam): Young & Rubicam, a marketing communications agency, has developed the Brand Asset Valuator, a tool to diagnose the power
and value of a brand. In using it, the agency surveys consumers’ perspectives along four
dimensions:
■
■
■
■
Differentiation: The defining characteristics of the brand and its distinctiveness
relative to competitors.
Relevance: The appropriateness and connection of the brand to a given
consumer.
Esteem: Consumers’ respect for and attraction to the brand.
Knowledge: Consumers’ awareness of the brand and understanding of what it
represents.
Chapter 4 Product and Portfolio Management
141
Young & Rubicam maintains that these criteria reveal important factors behind brand
strength and market dynamics. For example, although powerful brands score high on
all four dimensions, growing brands may earn higher grades for Differentiation and
Relevance, relative to Knowledge and Esteem. Fading brands often show the reverse
pattern, as they’re widely known and respected but may be declining toward commoditization or irrelevance (see Figure 4.4).
Growth Brand
Di
ffe
r
Re
Es
Kn
te
o
l
en e v a n em w l e
dg
tia
c
e
tio e
n
Strong Brand
Di
ffe
r
Re
Es
Kn
te
o
l
en e v a n em w l e
dg
tia
c
e
tio e
n
Declining Brand
Weak Brand
Di
Figure 4.4
ffe
r
Es
Re
Kn
te
o
l
en e v a n em w l e
dg
tia
c
e
e
tio
n
Di
ffe
r
Es
Re
Kn
te
o
l
en evan em w l e
dg
tia
c
e
e
tio
n
Young & Rubicam Brand Asset Valuator Patterns of Brand Equity
The Brand Asset Valuator is a proprietary tool, but the concepts behind it have broad
appeal. Many marketers apply these concepts by conducting independent research
and exercising judgment about their own brands relative to the competition. Leon
Ramsellar3 of Philips Consumer Electronics, for example, has reported using four key
measures in evaluating brand equity and offered sample questions for assessing them.
■
Uniqueness: Does this product offer something new to me?
■
Relevance: Is this product relevant for me?
■
Attractiveness: Do I want this product?
■
Credibility: Do I believe in the product?
142
MARKETING METRICS
Clearly Ramsellar’s list is not the same as Y&R’s BAV, but the similarity of the first two
factors is hard to miss.
Brand Finance: Founded in 1996, Brand Finance is an independent intangible asset
valuation consultancy based in the UK with offices in more than 20 countries. They
focus on valuation of the firm’s intangible assets.
Their methodology specifies three alternative brand valuation approaches: the Market,
Cost, and Income approaches.
Royalty Relief Methodology:
Brand Finance uses a Royalty Relief methodology when calculating the value of a brand.
This approach is based on the assumption that if the company did not own the trademarks that it benefits from, it would need to license them from a third party brand
owner and pay a licensing fee. Ownership of the trademarks relieves the company from
paying this royalty fee.
Royalty Relief Valuation Formula:
The Royalty Relief method involves estimating likely future sales, applying an appropriate royalty rate to them, and then discounting estimated future, post-tax royalties, to
arrive at a Net Present Value. This is held to represent the brand value.
Brand Ratings:
In determining the appropriate royalty rate, Brand Finance establishes a range of comparable royalty rates, and determines the point within the range that the brand under
review falls by reference to a brand rating. The brand rating is calculated using Brand
Finance’s BrandBeta analysis, which benchmarks the strength, risk and potential of a
brand, relative to its competitors, on a scale ranging from AAA to D. It is conceptually similar to a credit rating. The data used to calculate the ratings comes from various sources including Bloomberg, annual reports, client-commissioned research, and
Brand Finance internal research.
BrandZ (Millward Brown): BrandZ is Millward Brown’s brand equity database. It
holds data from more than two million consumers and professionals across 31 countries, comparing more than 23,000 brands. The database is used to estimate brand valuations, and each year since 2006, has been used to generate a list of the top 100 global
brands.
Its methodology’s objective is to peel away all financial components of brand value to
determine how much brand alone contributes to brand value. BrandZ conducts worldwide, ongoing, in-depth quantitative consumer research, and builds up a global picture
of brands on a category-by-category, and country-by-country basis. BrandZ uses the
following pillars to anchor its brand valuation: Meaningful, Different, and Salient.
Chapter 4 Product and Portfolio Management
143
Meaningful: In any category, these brands appeal more, generate greater “love,”
and meet the individual’s expectations and needs.
Different: These brands are unique in a positive way and “set the trends,” staying ahead of the curve for the benefit of the consumer.
Salient: They come spontaneously to mind as the brand of choice for key needs.
Valuation Process:
Step 1: Calculating Financial Value
Part A
■
■
■
Determine total corporate earnings from the corporation’s entire portfolio of
brands
Analyze financial information from annual reports and other sources, such as
Kantar Worldpanel and Kantar Retail. This analysis yields a metric called the
Attribution Rate.
Multiply Corporate Earnings by the Attribution Rate to arrive at Branded
Earnings, the amount of Corporate Earnings attributed to a particular brand.
If the Attribution Rate of a brand is 50 percent, for example, then half the
Corporate Earnings are identified as coming from that brand.
Part B
■
Determine future earnings and attribute an earnings multiple to the company. Information supplied by Bloomberg data is used to calculate the Brand
Multiple. BrandZ takes the Branded Earnings and multiplies that number by
the Brand Multiple to arrive at the Financial Value.
Step 2: Calculating Brand Contribution
■
■
144
We now have the value of the branded business as a proportion of the total
value of the corporation. But this branded business value is still not quite the
core that we are after. To arrive at Brand Value, we need to peel away a few
more layers, such as other factors that influence the value of the branded business; for example: price, convenience, availability, and distribution.
Because a brand exists in the mind of the consumer, we have to assess the
brand’s uniqueness and its ability to stand out from the crowd, generate
desire, and cultivate loyalty. This unique role played by brand is called Brand
Contribution.
MARKETING METRICS
Step 3: Calculating Brand Value
■
Take the Financial Value and multiply it by Brand Contribution, which is
expressed as a percentage of Financial Value. The result is Brand Value. Brand
Value is the dollar amount a brand contributes to the overall value of a corporation. Isolating and measuring this intangible asset reveals an additional
source of shareholder value that otherwise would not exist.
Brand Valuation Model (Interbrand): Interbrand, a division of Omnicom, is a brand
consultancy headquartered in New York City.
Interbrand publishes the Best Global Brands report on an annual basis. The report
identifies the world’s 100 most valuable brands. Interbrand’s methodology is also the
first of its kind to be ISO certified. To develop the report, Interbrand examines three
key aspects that contribute to a brand’s value:
■
The financial performance of the branded products or service.
■
The role the brand plays in influencing consumer choice.
■
The strength the brand has to command a premium price, or secure earnings
for the company.
Interbrand has refined its brand valuation into a five-step Economic Value Added
methodology. Through a similar methodology, Interbrand releases an annual ranking
of the best global brands called “Best Global Brands,” which evaluates each brand’s
financial performance, role, and strength. To qualify, brands must have a presence on
at least three major continents, and must have broad geographic coverage in growing
and emerging markets. Thirty percent of revenues must come from outside the home
country, and no more than fifty percent of revenues should come from any one continent. Economic profit must be expected to be positive over the longer term, delivering
a return above the brand’s cost of capital. The brand must have a public profile and
awareness across the major economies of the world.
Three key components to brand valuation methodology:
1. Financial Analysis—This measures the overall financial return to an organization’s investors, or its “economic profit.” Economic profit is the after-tax operating profit of the brand minus a charge for the capital used to generate the
brand’s revenue and margins.
2. Role of Brand—Role of Brand measures the portion of the purchase decision
attributable to the brand, as opposed to other factors (for example, purchase
drivers like price, convenience, or product features). The Role of Brand Index
(RBI) quantifies this as a percentage. RBI determinations for Best Global
Chapter 4 Product and Portfolio Management
145
Brands derive, depending on the brand, from one of three methods: primary
research, a review of historical roles of brands for companies in that industry,
or expert panel assessment.
3. Brand Strength—Brand Strength measures the ability of the brand to create loyalty and, therefore, sustainable demand and profit into the future.
Brand Strength analysis is based on an evaluation across several factors that
Interbrand believes makes a strong brand (factors are weighted and add up
to 100). These factors include leadership (25), stability (15), market (10), geographic spread (25), trend (10), support (10), and protection (5). Performance
on these factors is judged relative to other brands in the industry and relative
to other world-class brands. The Brand Strength analysis delivers a snapshot of
the strengths and weaknesses of the brand and is used to generate a road map
of activity to enhance the strength and value of the brand in the future.
Conjoint Analysis: Marketers use conjoint analysis to measure consumers’ preference
for various attributes of a product, service, or provider, such as features, design, price,
or location (see Section 4.5). By including brand and price as two of the attributes under
consideration, they can gain insight into consumers’ valuation of a brand—that is, their
willingness to pay a premium for it.
Data Sources, Complications, and Cautions
The methods described previously represent experts’ best attempts to place a value on
a complex and intangible entity. Almost all of the metrics in this book are relevant to
brand equity along one dimension or another.
Related Metrics and Concepts
Brand strategy is a broad field and includes several concepts that at first may appear to
be measurable. Strictly speaking, however, brand strategy is not a metric.
Brand Identity: This is the marketer’s vision of an ideal brand—the company’s goal for
perception of that brand by its target market. All physical, emotional, visual, and verbal
messages should be directed toward realization of that goal, including name, logo, signature, and other marketing communications. Brand Identity, however, is not stated in
quantifiable terms.
Brand Position and Brand Image: These refer to consumers’ actual perceptions of a
brand, often relative to its competition. Brand Position is frequently measured along
product dimensions that can be mapped in multi-dimensional space. If measured consistently over time, these dimensions may be viewed as metrics—as coordinates on a
perceptual map. (See Section 2.7 for a discussion of attitude, usage measures, and the
hierarchy of effects.)
146
MARKETING METRICS
Product Differentiation: This is one of the most frequently used terms in marketing, but it has no universally agreed-upon definition. More than mere “difference,” it
generally refers to distinctive attributes of a product that generate increased customer
preference or demand. These are often difficult to view quantitatively because they may
be actual or perceived, as well as non-monotonic. In other words, although certain
attributes such as price can be quantified and follow a linear preference model (that is,
either more or less is always better), others can’t be analyzed numerically or may fall
into a sweet spot, outside of which neither more nor less would be preferred (the spiciness of a food, for example). For all these reasons, Product Differentiation is hard to
analyze as a metric and has been criticized as a “meaningless term.”
Additional Citation
Simon, Julian, “Product Differentiation”: A Meaningless Term and an Impossible
Concept, Ethics, Vol. 79, No. 2 (Jan., 1969), pp. 131-138. Published by The University
of Chicago Press.
4.5 Conjoint Utilities and Consumer Preference
Conjoint utilities measure consumer preference for an attribute level and then—
by combining the valuations of multiple attributes—measure preference for an
overall choice. Measures are generally made on an individual basis, although this
analysis can also be performed on a segment level. In the frozen pizza market, for
example, conjoint utilities can be used to determine how much a customer values
superior taste (one attribute) versus paying extra for premium cheese (a second
attribute).
Conjoint utilities can also play a role in analyzing compensatory and noncompensatory decisions. Weaknesses in compensatory factors can be made up in
other attributes. A weakness in a non-compensatory factor cannot be overcome by
other strengths.
Conjoint analysis can be useful in determining what customers really want and—
when price is included as an attribute—what they’ll pay for it. In launching new
products, marketers find such analyses useful for achieving a deeper understanding of the values that customers place on various product attributes. Throughout
product management, conjoint utilities can help marketers focus their efforts on
the attributes of greatest importance to customers.
Chapter 4 Product and Portfolio Management
147
Purpose: To understand what customers want.
Conjoint analysis is a method used to estimate customers’ preferences, based on how
customers weight the attributes on which a choice is made. The premise of conjoint
analysis is that a customer’s preference between product options can be broken into
a set of attributes that are weighted to form an overall evaluation. Rather than asking
people directly what they want and why, in conjoint analysis, marketers ask people
about their overall preferences for a set of choices described on their attributes and
then decompose those into the component dimensions and weights underlying them.
A model can be developed to compare sets of attributes to determine which represents
the most appealing bundle of attributes for customers.
Conjoint analysis is a technique commonly used to assess the attributes of a product or
service that are important to targeted customers and to assist in the following:
■
Product design
■
Advertising copy
■
Pricing
■
Segmentation
■
Forecasting
Construction
Conjoint Analysis: A method of estimating customers by assessing the overall
preferences customers assign to alternative choices.
An individual’s preference can be expressed as the total of his or her baseline preferences for any choice, plus the partworths (relative values) for that choice expressed by
the individual.
In linear form, this can be represented by the following formula:
Conjoint Preference Linear Form (I) = [Partworth of Attribute1 to Individual (I)
* Attribute Level (1)] + [Partworth of
Attribute2 to Individual (I) * Attribute Level
(2)] + [Partworth of Attribute3 to Individual
(I) * Attribute Level (3)] + etc.
EXAMPLE: Two attributes of a cell phone, its price and its size, are ranked through conjoint
analysis, yielding the results shown in Table 4.5.
148
MARKETING METRICS
This could be read as follows:
Table 4.5 Conjoint Analysis: Price and Size of a Cell Phone
Attribute
Level
Partworth
Price
$100
0.9
Price
$200
0.1
Price
$300
–1
Size
Small
0.7
Size
Medium
–0.1
Size
Large
–0.6
A small phone for $100 has a partworth to customers of 1.6 (derived as 0.9 + 0.7). This is the highest
result observed in this exercise. A small but expensive ($300) phone is rated as –0.3 (that is, –1 +
0.7). The desirability of this small phone is offset by its price. A large, expensive phone is least desirable to customers, generating a partworth of –1.6 (that is, –1 + –0.6).
On this basis, we determine that the customer whose views are analyzed here would prefer a
medium-size phone at $200 (utility = 0) to a small phone at $300 (utility = –0.3). Such information
would be instrumental to decisions concerning the trade-offs between product design and price.
This analysis also demonstrates that, within the ranges examined, price is more important than
size from the perspective of this consumer. Price generates a range of effects from 0.9 to –1 (that is,
a total spread of 1.9), while the effects generated by the most and least desirable sizes span a range
only from 0.7 to –0.6 (total spread = 1.3).
Compensatory Versus Non-compensatory Consumer Decisions
A compensatory decision process is one in which a customer evaluates choices with
the perspective that strengths along one or more dimensions can compensate for weaknesses along others.
In a non-compensatory decision process, by contrast, if certain attributes of a product
are weak, no compensation is possible, even if the product possesses strengths along
other dimensions. In the previous cell phone example, for instance, some customers may feel that if a phone were greater than a certain size, no price would make it
attractive.
In another example, most people choose a grocery store on the basis of proximity. Any
store within a certain radius of home or work may be considered. Beyond that distance,
Chapter 4 Product and Portfolio Management
149
however, all stores will be excluded from consideration, and there is nothing a store can
do to overcome this. Even if it posts extraordinarily low prices, offers a stunningly wide
assortment, creates great displays, and stocks the freshest foods, for example, a store
will not entice consumers to travel 400 miles to buy their groceries.
Although this example is extreme to the point of absurdity, it illustrates an important
point: When consumers make a choice on a non-compensatory basis, marketers need
to define the dimensions along which certain attributes must be delivered, simply to
qualify for consideration of their overall offering.
One form of non-compensatory decision-making is elimination-by-aspect. In this
approach, consumers look at an entire set of choices and then eliminate those that do
not meet their expectations in the order of the importance of the attributes. In the selection of a grocery store, for example, this process might run as follows:
■
Which stores are within 5 miles of my home?
■
Which ones are open after 8 p.m.?
■
Which carry the spicy mustard that I like?
■
Which carry fresh flowers?
The process continues until only one choice is left.
In the ideal situation, in analyzing customers’ decision processes, marketers would have
access to information on an individual level, revealing
■
Whether the decision for each customer is compensatory or not
■
The priority order of the attributes
■
The “cut-off” levels for each attribute
■
The relative importance weight of each attribute if the decision follows a compensatory process
More frequently, however, marketers have access only to past behavior, helping them
make inferences regarding these items.
In the absence of detailed, individual information for customers throughout a market,
conjoint analysis provides a means to gain insight into the decision-making processes
of a sampling of customers. In conjoint analysis, we generally assume a compensatory
process. That is, we assume utilities are additive. Under this assumption, if a choice is
weak along one dimension (for example, if a store does not carry spicy mustard), it can
compensate for this with strength along another (for example, it does carry fresh-cut
flowers) at least in part. Conjoint analyses can approximate a non-compensatory model
by assigning non-linear weighting to an attribute across certain levels of its value.
For example, the weightings for distance to a grocery store might run as follows:
150
MARKETING METRICS
Within 1 mile:
0.9
1-5 miles away:
0.8
5-10 miles away:
–0.8
More than 10 miles away:
–0.9
In this example, stores outside a 5-mile radius cannot practically make up the loss of
utility they incur as a result of distance. Distance becomes, in effect, a non-compensatory
dimension.
By studying customers’ decision-making processes, marketers gain insight into the
attributes needed to meet consumer expectations. They learn, for example, whether
certain attributes are compensatory or non-compensatory. A strong understanding of
customers’ valuation of different attributes also enables marketers to tailor products
and allocate resources effectively.
Several potential complications arise in considering compensatory versus noncompensatory decisions. Customers often don’t know whether an attribute is compensatory or not, and they may not be readily able to explain their decisions. Therefore, it
is often necessary either to infer a customer’s decision-making process or to determine
that process through an evaluation of choices, rather than a description of the process.
It is possible, however, to uncover non-compensatory elements through conjoint analysis. Any attribute for which the valuation spread is so high that it cannot practically be
made up by other features is, in effect, a non-compensatory attribute.
EXAMPLE: Among grocery stores, Juan prefers the Acme market because it’s close to his
home, despite the fact that Acme’s prices are generally higher than those at the local Shoprite store.
A third store, Vernon’s, is located in Juan’s apartment complex. But Juan avoids it because Vernon’s
doesn’t carry his favorite soda.
From this information, we know that Juan’s shopping choice is influenced by at least three factors:
price, distance from his home, and whether a store carries his favorite soda. In Juan’s decision process, price and distance seem to be compensating factors. He trades price for distance. Whether the
soda is stocked seems to be a non-compensatory factor. If a store doesn’t carry Juan’s favorite soda,
it will not win his business, regardless of how well it scores on price and location.
Data Sources, Complications, and Cautions
Prior to conducting a conjoint study, it is necessary to identify the attributes of importance to a customer. Focus groups are commonly used for this purpose. After attributes
and levels are determined, a typical approach to Conjoint Analysis is to use a fractional
Chapter 4 Product and Portfolio Management
151
factorial orthogonal design, which is a partial sample of all possible combinations of
attributes. This is to reduce the total number of choice evaluations required by the
respondent. With an orthogonal design, the attributes remain independent of one
another, and the test doesn’t weigh one attribute disproportionately to another.
There are multiple ways to gather data, but a straightforward approach would be to
present respondents with choices and to ask them to rate those choices according to
their preferences. These preferences then become the dependent variable in a regression, in which attribute levels serve as the independent variables, as in the previous
equation. Conjoint utilities constitute the weights determined to best capture the preference ratings provided by the respondent.
Often, certain attributes work in tandem to influence customer choice. For example, a
fast and sleek sports car may provide greater value to a customer than would be suggested by the sum of the fast and sleek attributes. Such relationships between attributes
are not captured by a simple conjoint model, unless one accounts for interactions.
Ideally, conjoint analysis is performed on an individual level because attributes can be
weighted differently across individuals. Marketers can also create a more balanced view
by performing the analysis across a sample of individuals. It is appropriate to perform
the analysis within consumer segments that have similar weights. Conjoint analysis can
be viewed as a snapshot in time of a customer’s desires. It will not necessarily translate
indefinitely into the future.
It is vital to use the correct attributes in any conjoint study. People can only tell you their
preferences within the parameters you set. If the correct attributes are not included in a
study, while it may be possible to determine the relative importance of those attributes
that are included, and it may technically be possible to form segments on the basis of
the resulting data, the analytic results may not be valid for forming useful segments. For
example, in a conjoint analysis of consumer preferences regarding colors and styles of
cars, one may correctly group customers as to their feelings about these attributes. But if
consumers really care most about engine size, then those segmentations will be of little
value.
4.6 Segmentation Using Conjoint Utilities
Understanding customers’ desires is a vital goal of marketing. Segmenting, or
clustering similar customers into groups, can help managers recognize useful
patterns and identify attractive subsets within a larger market. With that understanding, managers can select target markets, develop appropriate offerings for
each, determine the most effective ways to reach the targeted segments, and allocate resources accordingly. Conjoint analysis can be highly useful in this exercise.
152
MARKETING METRICS
Purpose: To identify segments based on conjoint utilities.
As described in the previous section, conjoint analysis is used to determine customers’
preferences on the basis of the attribute weightings that they reveal in their decisionmaking processes. These weights, or utilities, are generally evaluated on an individual
level.
Segmentation entails the grouping of customers who demonstrate similar patterns of
preference and weighting with regard to certain product attributes, distinct from the
patterns exhibited by other groups. Using segmentation, a company can decide which
group(s) to target and can determine an approach to appeal to the segment’s members.
After segments have been formed, a company can set strategy based on their attractiveness (size, growth, purchase rate, diversity) and on its own capability to serve these
segments, relative to competitors.
Construction
To complete a segmentation based on conjoint utilities, one must first determine utility scores at an individual customer level. Next, one must cluster these customers into
segments of like-minded individuals. This is generally done through a methodology
known as cluster analysis.
Cluster Analysis: A technique that calculates the distances between customers
and forms groups by minimizing the differences within each group and maximizing the differences between groups.
Cluster analysis operates by calculating a “distance” (a sum of squares) between individuals and, in a hierarchical fashion, starts pairing those individuals together. The
process of pairing minimizes the “distance” within a group and creates a manageable
number of segments within a larger population.
EXAMPLE: The Samson-Finn Company has three customers. In order to help manage its marketing efforts, Samson-Finn wants to organize like-minded customers into segments. Toward that
end, it performs a conjoint analysis in which it measures its customers’ preferences among products
that are either reliable or very reliable, either fast or very fast (see Table 4.6). It then considers the
conjoint utilities of each of its customers to see which of them demonstrate similar wants. When
clustering on conjoint data, the distances would be calculated on the partworths.
Chapter 4 Product and Portfolio Management
153
Table 4.6
Customer Conjoint Utilities
Very Reliable
Reliable
Very Fast
Fast
Bob
0.4
0.3
0.6
0.2
Erin
0.9
0.1
0.2
0.7
Yogesh
0.3
0.3
0.5
0.2
The analysis looks at the difference between Bob’s view and Erin’s view on the importance of reliability on their choice. Bob’s score is 0.4 and Erin’s is 0.9. We can square the difference between
these to derive the “distance” between Bob and Erin.
Using this methodology, the distance between each pair of Samson-Finn’s customers can be calculated as follows:
Distances
Very Reliable
Reliable
Very Fast
Fast
Bob and Erin:
= (0.4 – 0.9)2
+ (0.3 – 0.1)2
+ (0.6 – 0.2)2
+ (0.2– 0.7)2
= 0.25
+ 0.04
+ 0.16
+ 0.25
+ (0.3 – 0.3)2
+ (0.6 – 0.5)2
+ (0.2 – 0.2)2
+ 0.0
+ 0.01
+ 0.0
+ (0.1 – 0.3)2
+ (0.2 – 0.5)2
+ (0.7 + 0.2)2
+ 0.04
+ 0.09
+ 0.25
= 0.7
Bob and Yogesh: = (0.4 – 0.3)2
= 0.01
= 0.02
Erin and Yogesh: = (0.9 – 0.3)2
= 0.36
= 0.74
On this basis, Bob and Yogesh appear to be very close to each other because their sum of squares
is 0.02. As a result, they should be considered part of the same segment. Conversely, in light of the
high sum-of-squares distance established by her preferences, Erin should not be considered a part
of the same segment with either Bob or Yogesh.
Of course, most segmentation analyses are performed on large customer bases. This example
merely illustrates the process involved in the cluster analysis calculations.
Data Sources, Complications, and Cautions
As noted previously, a customer’s utilities may not be stable, and the segment to
which a customer belongs can shift over time or across occasions. An individual might
belong to one segment for personal air travel, in which price might be a major factor,
and another for business travel, in which convenience might become more important.
154
MARKETING METRICS
Such a customer’s conjoint weights (utilities) would differ depending on the purchase
occasion.
Determining the appropriate number of segments for an analysis can be somewhat
arbitrary. There is no generally accepted statistical means for determining the “correct”
number of segments. Ideally, marketers look for a segment structure that fulfills the
following qualifications:
■
■
Each segment constitutes a homogeneous group, within which there is relatively little variance between attribute utilities of different individuals.
Groupings are heterogeneous across segments; that is, there is a wide variance
of attribute utilities between segments.
4.7 Conjoint Utilities and Volume Projection
The conjoint utilities of products and services can be used to forecast the market
share that each will achieve and the volume that each will sell. Marketers can
project market share for a given product or service on the basis of the proportion
of individuals who select it from a relevant choice set, as well as its overall utility.
Purpose: To use conjoint analysis to project the market share
and the sales volume that will be achieved by a product or service.
Conjoint analysis is used to measure the utilities for a product. The combination of these
utilities, generally additive, represents a scoring of sorts for the expected popularity of
that product. These scores can be used to rank products. However, further information
is needed to estimate market share. One can anticipate that the top-ranked product
in a selection set will have a greater probability of being chosen by an individual than
products ranked lower for that individual. Adding the number of customers who rank
the brand first should allow the calculation of customer share.
Data Sources, Complications, and Cautions
To complete a sales volume projection, it is necessary to have a full conjoint analysis.
This analysis must include all the important features according to which consumers
make their choice. Defining the “market” is clearly crucial to a meaningful result.
To define a market, it is important to identify all the choices in that market. Calculating
the percentage of “first choice” selections for each alternative merely provides a “share
Chapter 4 Product and Portfolio Management
155
of preferences.” To extend this to market share, one must estimate (1) the volume of
sales per customer, (2) the level of distribution or availability for each choice, and (3)
the percentage of customers who will defer their purchase until they can find their first
choice.
The greatest potential error in this process would be to exclude meaningful attributes
from the conjoint analysis.
Network effects can also distort a conjoint analysis. In some instances, customers do
not make purchase decisions purely on the basis of a product’s attributes but are also
affected by its level of acceptance in the marketplace. Such network effects, and the
importance of harnessing or overcoming them, are especially evident during shifts in
technology industries.
References and Suggested Further Reading
Aaker, D.A. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name,
New York: Free Press; Toronto; New York: Maxwell Macmillan; Canada: Maxwell Macmillan
International.
Aaker, D.A. (1996). Building Strong Brands, New York: Free Press.
Aaker, D.A., and J.M. Carman. (1982). “Are You Overadvertising?” Journal of Advertising
Research, 22(4), 57–70.
Aaker, D.A., and K.L. Keller. (1990). “Consumer Evaluations of Brand Extensions,” Journal of
Marketing, 54(1), 27–41.
Ailawadi, Kusum, and Kevin Keller. (2004). “Understanding Retail Branding: Conceptual
Insights and Research Priorities,” Journal of Retailing, Vol. 80, Issue 4, Winter, 331–342.
Ailawadi, Kusum, Donald Lehman, and Scott Neslin. (2003). “Revenue Premium As an
Outcome Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1–17.
Burno, Hernan A., Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing Face of
Measurement Tools Across the Product Lifecycle,” Does Marketing Measure Up? Performance
Metrics: Practices and Impact, Marketing Science Institute, No. 05-301.
Harvard Business School Case: Nestlé Refrigerated Foods Contadina Pasta & Pizza (A)
9-595-035. Rev Jan 30 1997.
Moran, Bill. Personal communication with Paul Farris.
156
MARKETING METRICS
5
CUSTOMER PROFITABILITY
Introduction
Key concepts covered in this chapter:
Customers, Recency, and Retention
Prospect Value Versus Customer Value
Customer Profit
Acquisition Versus Retention Spending
Customer Lifetime Value
Chapter 2, “Share of Hearts, Minds, and Markets,” presented metrics designed to measure
how well the firm is doing with its customers as a whole. Previously discussed metrics
were summaries of firm performance with respect to customers for entire markets or
market segments. In this chapter, we cover metrics that measure the performance of
individual customer relationships. We start with metrics designed to simply count how
many customers the firm serves. As this chapter will illustrate, it is far easier to count
the number of units sold than to count the number of people or businesses buying those
units. Section 5.2 introduces the concept of customer profit. Just as some brands are
more profitable than others, so too are some customer relationships. Whereas customer
profit is a metric that summarizes the past financial performance of a customer relationship, customer lifetime value looks forward in an attempt to value existing customer
relationships. Section 5.3 discusses how to calculate and interpret customer lifetime
value. One of the more important uses of customer lifetime value is to inform prospecting decisions. Section 5.4 explains how this can be accomplished and draws the careful
distinction between prospect and customer value. Section 5.5 discusses acquisition and
retention spending—two metrics firms track in order to monitor the performance of
these two important kinds of marketing spending—spending designed to acquire new
customers and spending designed to retain and profit from existing customers.
157
Metric
Construction
Considerations
5.1
Customers
The number of
people (businesses)
who bought from
the firm during
a specified time
period.
Avoid double
Measure how well the
counting people
firm is attracting and
who bought
retaining customers.
more than one
product. Carefully
define customer
as individual/
household/ screenname/ division who
bought/ordered/
registered.
5.1
Recency
The length of time
since a customer’s
last purchase.
In noncontractual
situations, the
firm will want to
track the recency
of its customers.
Track changes in
number of active
customers.
5.1
Retention Rate
The ratio of
customers
retained to the
number at risk.
Not to be
confused with
growth (decline)
in customer
counts. Retention
refers only to
existing customers
in contractual
situations.
Track changes in the
ability of the firm to
retain customers.
5.2
Customer Profit
The difference
between the
revenues earned
from and the
costs associated
with the customer
relationship
during a specified
period.
Requires assigning
revenues and
costs to individual
customers.
Allows the firm
to identify which
customers are
profitable and
which are not as
a precursor to
differential treatment
designed to improve
firm profitability.
158
MARKETING METRICS
Purpose
Metric
Construction
Considerations
Purpose
5.3
Customer Lifetime The present value
Value
of the future cash
flows attributed
to the customer
relationship.
Requires a
projection of
future cash flows
from a customer
relationship.
This will be
easier to do in
a contractual
situation.
Formulations of
CLV differ with
respect to the
treatment of the
initial margin
and acquisition
spending.
Customer
relationship
management
decisions should
be made with
the objective of
improving CLV.
Acquisition
budgeting should
be based on CLV.
5.4
Prospect Lifetime
Value
The response rate
times the sum of
the initial margin
and the CLV of the
acquired customer
minus the cost of
the prospecting
effort.
There are a variety
of equivalent
ways to do the
calculations
necessary to
see whether a
prospecting effort
is worthwhile.
To guide the
firm’s prospecting
decisions.
Prospecting is
beneficial only
if the expected
prospect lifetime
value is positive.
5.5
Average
Acquisition Cost
The ratio of
acquisition
spending to
the number of
new customers
acquired.
It is often
difficult to isolate
acquisition
spending from
total marketing
spending.
To track the cost
of acquiring new
customers and to
compare that cost
to the value of the
newly acquired
customers.
5.5
Average Retention
Cost
The ratio of
retention spending
to the number
of customers
retained.
It is often difficult
to isolate retention
spending from
total marketing
spending. The
average retention
cost number is not
very useful to help
make retention
budgeting
decisions.
To monitor
retention spending
on a per-customer
basis.
Chapter 5 Customer Profitability
159
5.1 Customers, Recency, and Retention
These three metrics are used to count customers and track customer activity irrespective of the number of transactions (or dollar value of those transactions) made
by each customer.
A customer is a person or business that buys from the firm.
■
■
■
Customer Counts: These are the number of customers of a firm for a specified time
period.
Recency: This refers to the length of time since a customer’s last purchase. A sixmonth customer is someone who purchased from the firm at least once within the
last six months.
Retention Rate: This is the ratio of the number of retained customers to the number
at risk.
In contractual situations, it makes sense to talk about the number of customers currently under contract and the percentage retained when the contract period runs out.
In non-contractual situations (such as catalog sales), it makes less sense to talk
about the current number of customers, but instead to count the number of customers of a specified recency.
Purpose: To monitor firm performance in attracting and
retaining customers.
Only recently have most marketers worried about developing metrics that focus on
individual customers. In order to begin to think about managing individual customer
relationships, the firm must first be able to count its customers. Although consistency
in counting customers is probably more important than formulating a precise definition, a definition is needed nonetheless. In particular, we think the definition of and
the counting of customers will be different in contractual versus non-contractual
situations.
Construction
Counting Customers
In contractual situations, it should be fairly easy to count how many customers are currently under contract at any point in time. For instance, Vodafone Australia,1 a global
mobile phone company, was able to report 2.6 million direct customers at the end of
the December quarter.
160
MARKETING METRICS
One complication in counting customers in contractual situations is the handling
of contracts that cover two or more individuals. Does a family plan that includes
five phones but one bill count as one or five? Does a business-to-business contract with
one base fee and charges for each of 1,000 phones in use count as one or 1,000 customers? Does the answer to the previous question depend on whether the individual users
pay Vodafone, pay their company, or pay nothing? In situations such as these, the firm
must select some standard definition of a customer (policy holder, member) and implement it consistently.
A second complication in counting customers in contractual situations is the treatment
of customers with multiple contracts with a single firm. USAA, a global insurance and
diversified financial services association, provides insurance and financial services to
the U.S. military community and their families. Each customer is considered a member,
complete with a unique membership number. This allows USAA to know exactly how
many members it has at any time—more than ten million at the end of 2014—most of
whom avail themselves of a variety of member services.
For other financial services companies, however, counts are often listed separately
for each line of business. The 2014 annual report for State Farm Insurance, for
example, lists a total of 82 million policies and accounts with a pie chart showing the
percentage breakdown among auto, homeowners, life, annuities, and so on. Clearly
the 82 million is a count of policies and not customers. Presumably because some
customers use State Farm for auto, home, and life insurance, they get double and even
triple counted in the 82 million number. Because State Farm knows the names and
addresses of all their policyholders, it seems feasible that they could count how many
individual customers they serve. The fact that State Farm counts policies and not
customers suggests an emphasis on selling policies rather than managing customer
relationships.
Finally, we offer an example of a natural gas company that went out of its way to double
count customers—defining a customer to be “a consumer of natural gas distributed in
any one billing period at one location through one meter. An entity using gas at separate locations is considered a separate customer at each location.” For this natural gas
company, customers were synonymous with meters. This is probably a great way to
view things if your job is to install and service meters. It is not such a great way to view
things if your job is to market natural gas.
In non-contractual situations, the ability of the firm to count customers depends on
whether individual customers are identifiable. If customers are not identifiable, firms
can only count visits or transactions. Because Wal-Mart does not identify its shoppers, its customer counts are nothing more than the number of transactions that
go through the cash registers in a day, week, or year. These “traffic” counts are akin
to turnstile numbers at sporting events and visits to a Web site. In one sense they
Chapter 5 Customer Profitability
161
count people, but when summed over several periods, they no longer measure separate individuals. So whereas home attendance at Atlanta Braves games in 19932 was
3,884,720, the number of people attending one or more Braves games that year was
some smaller number.
In non-contractual situations with identifiable customers (direct mail, retailers with
frequent shopper cards, warehouse clubs, purchases of rental cars and lodging that
requires registration), a complication is that customer purchase activity is sporadic.
Whereas The New York Times knows exactly how many current customers (subscribers) it has, the sporadic buying of cataloger L.L.Bean’s customers means that it makes
no sense to talk about the number of current L.L.Bean customers. L.L.Bean will know
the number of orders it receives daily, it will know the number of catalogs it mails
monthly, but it cannot be expected to know the number of current customers it has
because it is difficult to define a “current” customer.
Instead, firms in non-contractual situations count how many customers have bought
within a certain period of time. This is the concept of recency —the length of time since
the last purchase. Customers of recency one year or less are customers who bought
within the last year. Firms in non-contractual situations with identifiable customers will
count customers of various recencies.
Recency: The length of time since a customer’s last purchase.
For example, eBay reported 157 million active users in the second quarter of 2015.
Active users were defined as the number of users of the eBay platform who bid, bought,
or listed an item within the previous 12-month period. They go on to report that 149
million active users were reported in the same period a year ago.
Notice that eBay counts “active users” rather than “customers” and uses the concept of recency to track its number of active users across time. The number of active
(12-month) users increased from 149 million to 157 million in one year. This tells
the firm that the number of active customers increased due in part to customer acquisition. A measure of how well the firm maintained existing customer relationships is
the percentage of the 149 million active customers one year ago who were active in the
previous 12 months. That ratio measure is similar to retention in that it reflects the
percentage of active customers who remained active in the subsequent period.
Retention: Applies to contractual situations in which customers are either retained
or not. Customers either renew their magazine subscriptions or let them run out.
Customers maintain a checking account with a bank until they close it out. Renters
pay rent until they move out. These are examples of pure customer retention situations
where customers are either retained or considered lost for good.
162
MARKETING METRICS
In these situations, firms pay close attention to retention rates.
Retention Rate: The ratio of the number of customers retained to the number
at risk.
If 40,000 subscriptions to Fortune magazine are set to expire in July and the publisher
convinces 26,000 of those customers to renew, we would say that the publisher retained
65% of its subscribers.
The complement of retention is attrition or churn. The attrition or churn rate for the
40,000 Fortune subscribers was 35%.
Notice that this definition of retention is a ratio of the number retained to the number at risk (of not being retained). The key feature of this definition is that a customer
must be at risk of leaving in order to be counted as a customer successfully retained.
This means that new Fortune subscribers obtained during July are not part of the equation, nor are the large number of customers whose subscriptions were set to run out in
later months.
Finally, we point out that it sometimes makes better sense to measure retention in
“customer time” rather than “calendar time.” Rather than ask what the firm’s retention rate was in 2004, it may be more informative to ask what percentage of customers
surviving for three years were retained throughout year four.
Data Sources, Complications, and Cautions
The ratio of the total number of customers at the end of the period to the number of
customers at the beginning of the period is not a retention rate. Retention during the
period does affect this ratio, but customer acquisitions also affect the ratio.
The percentage of customers starting the period who remained customers throughout the period is a lot closer to being a retention rate. This percentage would be a true
retention rate if all the customers starting the period were at risk of leaving during
the period.
Advice on Counting Customers3
Defining the customer properly is critical.
Marketers tend to count “customers” in ways that are easy and consequently get
the wrong answers. They tend to gloss over the fundamental and critically important step of defining the customer. With the wrong definition, counting
doesn’t matter.
Chapter 5 Customer Profitability
163
Banks look at “households” because they are “relationship” obsessed (relationship
being defined as the number of products sold to customers with a common account
address). Banks tend to emphasize the number of products sold. No matter that
the household may contain a business owner with nearly all the accounts, a spouse
who banks mostly elsewhere, and children who do not bank at all. Household in
this situation is meaningless. There are at least three “customers” here: business
owner (a great customer), spouse (almost a non-customer), and kids (definitely
non-customers).
Retailers count transactions or “tickets” (cash register receipts), which may cover
stuff sold to Mom, Dad, and the kids, along with Aunt Mary and neighbor Sue. Or,
it may reflect a purchase by a spouse who is buying for his or her partner under
specific instructions. In this circumstance, the spouse giving the instructions is the
real customer, while the other merely implements the transaction.
Defining the customer is nearly always hard because it requires a clear understanding of both business strategy and buyer behavior.
Not all “customers” are the same.
Attracting and retaining “customers” cannot be measured for management action
purposes without understanding the differences between customers. Last year, a
major software firm we will call Zapp bought a single copy of a piece of software.
Another company we will call Tancat bought 100 copies. Are these both “customers?” Of course not. Tancat is almost certainly a customer that needs to be retained
and possibly expanded into other products. Zapp is probably just evaluating the
product in order to stay on top of new software concepts and potentially copy it.
One option is to follow up with Zapp with their one-copy purchase to see what is
really going on. Zapp could become a great “customer” if we understand what motivated their purchase or if we use that purchase to gain a contact base.
Before you count anything, you have to segment your potential and current product
or service users into groups that can be strategically addressed. Some current buyers
like Zapp are actually potential buyers in terms of what you should do about them.
You must count buyers and prospects who are alike in defined ways.
Where is the “customer?”
Large customers often buy independently from each user location. Is Bank of
America the customer, or is each branch office a customer? If Citicorp were to buy
centrally, how could you count it as one customer while Bank of America counts as
hundreds of customers?
Who is the “customer?”
Defining who is the customer is even trickier. Many “customers” are not those who
place the order with your salespeople. The real customer is deep within the bowels
164
MARKETING METRICS
of the buyer organization, someone whom it will take a great deal of effort to even
identify. The account name may be GM, but the real customer may be Burt Cipher,
an engineer in some unknown facility. Or, the Ford buyer may have consolidated
orders from several individuals scattered across the country. In this case, Ford is not
the customer for anything but billing purposes. So, what do you count?
Even more common is the multi-headed customer. Buying decisions are made by
several people. Different people may be central to a decision at different times or
for different products. Big companies have sales teams dedicated to selling into such
buying groups. Although they may be counted as a single customer, the dynamics
of their buying decision is substantially more complicated than decisions made by a
single individual.
Apparel retailers who sell pre-teen clothing have at least two customers: Mom and
the pre-teen wearer. Do you count one or both as customers? Marketing might
want to treat each as a customer for deciding how to design and place ads. The store
might treat them both as a single customer or choose the pre-teen as their target.
The key takeaway is that customer definition for counting depends fundamentally on the purpose of the count. You may have to count the same “customer” in
different ways for different purposes. There is no universal customer definition.
5.2 Customer Profit
Customer profit (CP) is the profit the firm makes from serving a customer or customer group over a specified period of time.
Calculating customer profitability is an important step in understanding which
customer relationships are better than others. Often, the firm will find that some
customer relationships are unprofitable. The firm may be better off (more profitable) without these customers. At the other end, the firm will identify its most
profitable customers and be in a position to take steps to ensure the continuation
of these most profitable relationships.
Purpose: To identify the profitability of individual customers.
Companies commonly look at their performance in aggregate. A common phrase
within a company is something like: “We had a good year, and the business units
delivered $400,000 in profits.” When customers are considered, it is often using an
average such as “We made a profit of $2.50 per customer.” Although these can be
Chapter 5 Customer Profitability
165
useful metrics, they sometimes disguise an important fact that not all customers are
equal and, worse yet, some are unprofitable. Simply put, rather than measuring the
“average customer,” we can learn a lot by finding out what each customer contributes
to our bottom line.4
Customer Profitability: The difference between the revenues earned from and
the costs associated with the customer relationship during a specified period.
The overall profitability of the company can be improved by treating dissimilar
customers differently.
In essence, think of three different tiers of customer:
1. Top Tier customers—REWARD: Your most valuable customers are the ones
you most want to retain. They should receive more of your attention than any
other group. If you lose these guys, your profit suffers the most. Look to reward
them in ways other than simply lowering your price. These customers probably
value what you do the most and may not be price-sensitive.
2. Second Tier customers—GROW: The customers in the middle—with middle to
low profits associated with them—might be targeted for growth. Here you have
customers whom you may be able to develop into Top Tier customers. Look to
the share of customer metrics described in Section 5.3 to help figure out which
customers have the most growth potential.
3. Third Tier customers—FIRE: The company loses money on servicing these
people. If you cannot easily promote them to the higher tiers of profitability,
you should consider charging them more for the services they currently consume. If you can recognize this group beforehand, it may be best not to acquire
these customers in the first place.
A database that can analyze the profitability of customers at an individual level can be
a competitive advantage. If you can figure out profitability by customer, you have a
chance to defend your best customers and maybe even poach the most profitable consumers from your competitors.
Construction
In theory, this is a trouble-free calculation. Find out the cost to serve each customer and
the revenues associated with each customer for a given period. Do the subtraction to
get profit for the customer and sort the customers based on profit. Although painless in
theory, large companies with a multitude of customers will find this a major challenge
even with the most sophisticated of databases.
166
MARKETING METRICS
To do the analysis with large databases, it may be necessary to abandon the notion of
calculating profit for each individual customer and work with meaningful groups of
customers instead.
After you have the sorted list of customer profits (or customer-group profits), the custom is to plot cumulative percentage of total profits versus cumulative percentage of
total customers. Given that the customers are sorted from highest to lowest profit, the
resulting graph usually looks something like the head of a whale.
Profitability will increase sharply and tail off from the very beginning. (Remember,
our customers have been sorted from most to least profitable.) Whenever there are
some negative profit customers, the graph reaches a peak—above 100%—as profit per
customer moves from positive to negative. As we continue through the negative-profit
customers, cumulative profits decrease at an ever-increasing rate. The graph always
ends at 100% of the customers accounting for 100% of the total profit.
Robert Kaplan (co-developer of Activity-Based Costing and the Balanced Scorecard)
likes to refer to these curves as “whale curves.”5 In Kaplan’s experience, the whale
curve usually reveals that the most profitable 20% of customers can sometimes generate between 150% and 300% of total profits so that the resulting curve resembles
a sperm whale rising above the water’s surface. See Figure 5.2 for an example of a
whale curve.
EXAMPLE: A catalog retailer has grouped customers in 10 deciles based on profitability (see
Table 5.1 and Figure 5.1). (A decile is a tenth of the population, so 0-10% is the most profitable 10%
of customers.)
10–20%
20–30%
30–40%
40–50%
50–60%
60–70%
70–80%
80–90%
90–100%
Customer Profitability Ranked by Profitability
0–10%
Table 5.1
Band ($m)
Profitability
$100
$50
$25
$10
$5
$3
$2
$0
($8)
($20)
% of Total
Profits
60%
30%
15%
6%
3%
2%
1%
0%
–5%
Customers
Decile by
Profitability
–12%
Here we have a clear illustration that if they were no longer to serve the least profitable 20% of customers, they would be $28 million better off.
Chapter 5 Customer Profitability
167
Customer Profitability
70%
60%
% of Company Profits
50%
40%
30%
20%
10%
0%
0-10%
–10%
10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90% 90100%
–20%
Decile
Figure 5.1 Customer Profitability by Decile
Table 5.2 presents this same customer information in cumulative form. Cumulative profits plotted
across deciles begins to look like a whale with a steeply rising ridge reaching a peak of total profitability above 100% and tapering off thereafter (see Figure 5.2).
0–10%
10–20%
20–30%
30–40%
40–50%
50–60%
60–70%
70–80%
80–90%
90–100%
Table 5.2 Cumulative Profitability Peaks Before All Customers Are Served
Cumulative
Profits
$100
$150
$175
$185
$190
$193
$195
$195
$187
$167
Cumulative
Profits %
59.9
89.8
104.8
110.8
113.8
115.6
116.8
116.8
112.0 100.0
Customers
Decile by
Profitability
168
MARKETING METRICS
Cumulative Profits
140%
120%
Cumulative Profits
100%
80%
60%
40%
20%
%
00
–1
90
0%
–9
80
0%
–8
70
0%
–7
60
50
–6
0%
0%
–5
40
0%
–4
30
0%
–3
20
0%
–2
10
0–
10
%
0%
Decile
Figure 5.2
The Whale Curve
Data Sources, Complications, and Cautions
Measuring customer profitability requires detailed information. Assigning revenues to
customers is often the easy part; assigning your costs to customers is much harder. The
cost of goods sold obviously gets assigned to the customers based on the goods each
customer purchased. Assigning the more indirect costs may require the use of some
form of activity-based costing (ABC) system. Finally, there may be some categories of
costs that will be impossible to assign to the customer. If so, it is probably best to keep
these costs as company costs and be content with the customer profit numbers adding
up to something a little different to the total company profit.
When considering the profits from customers, it must be remembered that most things
change over time. Customers who were profitable last year may not be profitable
Chapter 5 Customer Profitability
169
this year. Because the whale curve reflects past performance, we must be careful when
using it to make decisions that shape the future. For example, we may very well want
to continue a relationship that was unprofitable in the past if we know things will
change for the better in the future. For example, banks typically offer discount packages to students to gain their business. This may well show low or negative customer
profits in the short term. The “plan” is that future profits will compensate for current
losses. Customer lifetime value (addressed in Section 5.3) is a forward-looking metric that attempts to account for the anticipated future profitability of each customer
relationship.
When capturing customer information to decide which customers to serve, it is important to consider the legal environment in which the company operates. This can change
considerably across countries, where there may be anti-discrimination laws and special
situations in some industries. For instance, public utilities are sometimes obligated to
serve all customers.
It is also worth remembering that intrusive capturing of customer-specific data can
damage customer relationships. Some individuals will be put off by excess data gathering. For a food company, it may help to know which of your customers are on a diet.
But the food company’s management should think twice before adding this question to
their next customer survey.
Sometimes there are sound financial reasons for continuing to serve unprofitable customers. For example, some companies rely on network effects. Take the case of the
United States Postal Service—part of its strength is the ability to deliver to the whole
country. It may superficially seem profitable to stop deliveries to remote areas. But
when that happens, the service becomes less valuable for all customers. In short, sometimes unprofitable customer relationships are necessary for the firm to maintain their
profitable ones.
Similarly, companies with high fixed costs that have been assigned to customers during the construction of customer profit must ask whether those costs will go away if
they terminate unprofitable customer relationships. If the costs do not go away, ending
unprofitable relationships may only serve to make the surviving relationships look even
less profitable (after the reallocation of costs) and result in the lowering of company
profits. In short, make certain that the negative profit goes away if the relationship is
terminated. Certainly the revenue and cost of goods sold will go away, but if some of the
other costs do not, the firm could be better off maintaining a negative profit relationship as it contributes to covering fixed cost (refer to Sections 3.4 and 3.6).
Abandoning customers is a very sensitive practice, and a business should always consider the public relations consequences of such actions. Similarly, when you get rid of a
customer, you cannot expect to attract them back very easily should they migrate into
your profitable segment.
170
MARKETING METRICS
Finally, because the whale curve examines cumulative percentage of total profits, the
numbers are very sensitive to the dollar amount of total profit. When the total dollar
profit is a small number, it is fairly easy for the most profitable customers to represent
a huge percentage of that small number. So when you hear that 20% of the firm’s customers represent 350% of the firm’s profit, one of the first things you should consider is
the total dollar value of profits. If that total is small, 350% of it can also be a fairly small
number of dollars. To cement this idea, ask yourself what the whale curve would look
like for a firm with $0 profit.
5.3 Customer Lifetime Value
Customer lifetime value is the dollar value of a customer relationship based on the
present value of the projected future cash flows from the customer relationship.
When margins and retention rates are constant, the following formula can be used
to calculate the lifetime value of a customer relationship:
Customer Lifetime Value ($) = Margin ($) *
Retention Rate (%)
1 + Discount Rate (%) – Retention Rate (%)
Customer lifetime value (CLV) is an important concept in that it encourages
firms to shift their focus from quarterly profits to the long-term health of their
customer relationships. Customer lifetime value is an important number because
it represents an upper limit on spending to acquire new customers.
Purpose: To assess the value of each customer.
As Don Peppers and Martha Rogers are fond of saying, “some customers are more
equal than others.”6 We saw a vivid illustration of this in the last section, which examined the profitability of individual customer relationships. As we noted, customer profit
(CP) is the difference between the revenues and the costs associated with the customer
relationship during a specified period. The central difference between CP and customer
lifetime value (CLV) is that CP measures the past and CLV looks forward. As such,
CLV can be more useful in shaping managers’ decisions but is much more difficult to
quantify. Quantifying CP is a matter of carefully reporting and summarizing the results
of past activity, whereas quantifying CLV involves forecasting future activity.
Customer Lifetime Value (CLV): The present value of the future cash flows
attributed to the customer relationship.
Chapter 5 Customer Profitability
171
The concept of present value will be talked about in more detail in Section 10.4. For
now, you can think of present value as the discounted sum of future cash flows. We discount (multiply by a carefully selected number less than one) future cash flows before
we add them together to account for the fact that there is a time value of money. The
time value of money is another way of saying that everyone would prefer to get paid
sooner rather than later and everyone would prefer to pay later rather than sooner. This
is true for individuals (the sooner I get paid, the sooner I can pay down my credit card
balance and avoid interest charges) as well as for firms. The exact discount factors used
depend on the discount rate chosen (10% per year as an example) and the number of
periods until we receive each cash flow (dollars received 10 years from now must be
discounted more than dollars received five years in the future).
The concept of CLV is nothing more than the concept of present value applied to
cash flows attributed to the customer relationship. Because the present value of any
stream of future cash flows is designed to measure the single lump sum value today of
the future stream of cash flows, CLV will represent the single lump sum value today
of the customer relationship. Even more simply, CLV is the dollar value of the customer relationship to the firm. It is an upper limit on what the firm would be willing
to pay to acquire the customer relationship as well as an upper limit on the amount
the firm would be willing to pay to avoid losing the customer relationship. If we view
a customer relationship as an asset of the firm, CLV would present the dollar value of
that asset.
Cohort and Incubate
One way to project the value of future customer cash flows is to make the heroic
assumption that the customers acquired several periods ago are no better or worse (in
terms of their CLV) than the ones we currently acquire. We then go back and collect
data on a cohort of customers all acquired at about the same time and carefully reconstruct their cash flows over some finite number of periods. The next step is to discount
the cash flows for each customer back to the time of acquisition to calculate that customer’s sample CLV and then average all of the sample CLVs together to produce an
estimate of the CLV of each newly acquired customer. We refer to this method as the
“cohort and incubate” approach. Equivalently, one can calculate the present value of
the total cash flows from the cohort and divide by the number of customers to get the
average CLV for the cohort. If the value of customer relationships is stable across time,
the average CLV of the cohort sample is an appropriate estimator of the CLV of newly
acquired customers.
As an example of this cohort and incubate approach, Berger, Weinberg, and Hanna
(2003) followed all the customers acquired by a cruise-ship line in 1993. The 6,094
customers in the cohort of 1993 were tracked (incubated) for five years. The total net
172
MARKETING METRICS
present value of the cash flows from these customers was $27,916,614. These flows
included revenues from the cruises taken (the 6,094 customers took 8,660 cruises
over the five-year horizon), variable cost of the cruises, and promotional costs. The
total five-year net present value of the cohort expressed on a per-customer basis came
out to be $27,916,614/6,094 or $4,581 per customer. This is the average five-year CLV
for the cohort.
“Prior to this analysis, [cruise-line] management would never spend more than
$3,314 to acquire a passenger . . . Now, aware of CLV (both the concept and the
actual numerical results), an advertisement that [resulted in a cost per acquisition of $3 to $4 thousand] was welcomed—especially because the CLV numbers are
conservative (again, as noted, the CLV does not include any residual business after
five years.)”7
The cohort and incubate approach works well when customer relationships are
stationary—changing slowly over time. When the value of relationships changes
slowly, we can use the value of incubated past relationships as predictive of the value of
new relationships.
In situations where the value of customer relationships changes more rapidly, firms
often use a simple model to forecast the value of those relationships. By a model,
we mean some assumptions about how the customer relationship will unfold. If the
model is simple enough, it may even be possible to find an equation for the present
value of our model of future cash flows. This makes the calculation of CLV even easier
because it now requires only the substitution of numbers for our situation into the
equation for CLV.
Next, we will explain what is perhaps the simplest model for future customer cash flows
and the equation for the present value of those cash flows. Although it’s not the only
model of future customer cash flows, this one gets used the most.
Construction
The model for customer cash flows treats the firm’s customer relationships as something of a leaky bucket. Each period, a fraction (1 less the retention rate) of the firm’s
customers leave and are lost for good.
The CLV model has only three parameters: 1) constant margin (contribution after
deducting variable costs including retention spending) per period, 2) constant retention probability per period, and 3) discount rate. Furthermore, the model assumes that
in the event that the customer is not retained, they are lost for good. Finally, the model
assumes that the first margin will be received (with probability equal to the retention
rate) at the end of the first period.
Chapter 5 Customer Profitability
173
The one other assumption of the model is that the firm uses an infinite horizon when it
calculates the present value of future cash flows. Although no firm actually has an infinite horizon, the consequences of assuming one are discussed in the following section.
Customer Lifetime Value: The CLV formula8 multiplies the per-period cash margin
(hereafter we will just use the term “margin”) by a factor that represents the present
value of the expected length of the customer relationship:
Customer Lifetime Value ($) = Margin ($) *
Retention Rate (%)
1 + Discount Rate (%) – Retention Rate (%)
Under the assumptions of the model, CLV is a multiple of the margin. The multiplicative factor represents the present value of the expected length (number of periods)
of the customer relationship. When retention equals 0, the customer will never be
retained, and the multiplicative factor is zero. When retention equals 1, the customer
is always retained, and the firm receives the margin in perpetuity. The present value of
the margin in perpetuity turns out to be Margin/Discount Rate. For retention values in
between, the CLV formula tells us the appropriate multiplier.
EXAMPLE: An Internet Service Provider (ISP) charges $19.95 per month. Variable costs are
about $1.50 per account per month. With marketing spending of $6 per year, their attrition is only
0.5% per month. At a monthly discount rate of 1%, what is the CLV of a customer?
Contribution Margin = ($19.95 – $1.50 – $6/12)= $17.95
Retention Rate = 0.995
Discount Rate = 0.01
Customer Lifetime Value (CLV) = Margin *
Retention Rate (%)
1 + Discount Rate (%) – Retention Rate (%)
CLV = $17.95 * [0.995/(1 + 0.01 – 0.995)]
CLV = [$17.95] * [66.33]
CLV = $1,191
Data Sources, Complications, and Cautions
The retention rate (and by extension the attrition rate) is a driver of customer value.
Very small changes can make a major difference to the lifetime value calculated.
Accuracy in this parameter is vital to meaningful results.
174
MARKETING METRICS
The retention rate is assumed to be constant across the life of the customer relationship.
For products and services that go through a trial, conversion, and loyalty progression,
retention rates will increase over the lifetime of the relationship. In those situations,
the model explained here might be too simple. If the firm wants to estimate a sequence
of retention rates, a spreadsheet model might be more useful in calculating CLV.
The discount rate is also a sensitive driver of the lifetime value calculation—as with
retention, seemingly small changes can make major differences to customer lifetime
value. The discount rate should be chosen with care.
The contribution is assumed to be constant across time. If margin is expected to increase
over the lifetime of the customer relationship, the simple model will not apply.
Take care not to use this CLV formula for relationships in which customer inactivity
does not signal the end of the relationship. In catalog sales, for example, a small percentage of the firm’s customers purchase from any given catalog. Don’t confuse the
percentage of customers active in a given period (relevant for the cataloger) with the
retention rates in this model. If customers often return to do business with the firm
after a period of inactivity, this CLV formula does not apply.
Customer Lifetime Value (CLV) with Initial Margin: One final source of confusion
concerns the timing assumptions inherent in the model. The first cash flow accounted
for in the model is the margin received at the end of one period with probability equal
to the retention rate. Other models also include an initial margin received at the beginning of the period. If a certain receipt of an initial margin is included, the new CLV will
equal the old CLV plus the initial margin. Furthermore, if the initial margin is equal to
all subsequent margins, there are at least two ways to write formulas for the CLV that
include the initial margin:
Retention Rate (%)
CLV with Initial = Margin ($) + Margin ($)
Margin ($)
1 + Discount Rate (%) – Retention Rate (%)
or
= Margin ($) *
1 + Discount Rate (%)
1 + Discount Rate (%) – Retention Rate (%)
The second formula looks just like the original formula with 1 + Discount Rate taking the place of the retention rate in the numerator of the multiplicative factor. Just
remember that the new CLV formula and the original CLV formula apply to the same
situations and differ only in the treatment of an initial margin. This new CLV formula
includes it, whereas the original CLV formula does not.
Chapter 5 Customer Profitability
175
The Infinite Horizon Assumption
In some industries and companies it is typical to calculate four- or five-year customer
values instead of using the infinite time horizon inherent in the previous formulas. Of
course, over shorter periods customer retention rates are less likely to be affected by
major shifts in technology or competitive strategies and are more likely to be captured
by historical retention rates. For managers, the question is “Does it make a difference
whether I use the infinite time horizon or (for example) the five-year customer value?”
The answer to this question is yes, sometimes, it can make a difference because the
value over five years can be less than 70% of the value over an infinite horizon (see
Table 5.3).
Table 5.3 calculates the percentages of (infinite horizon) CLV accruing in the first five
years. If retention rates are higher than 80% and discount rates are lower than 20%,
differences in the two approaches will be substantial. Depending on the strategic risks
that companies perceive, the additional complexities of using a finite horizon can be
informative.
Table 5.3 Finite-Horizon CLV As a Percentage of Infinite-Horizon CLV
Percent of CLV Accruing in First Five Years
Discount
Rates
Retention Rates
40%
50%
60%
70%
80%
90%
2%
99%
97%
93%
85%
70%
47%
4%
99%
97%
94%
86%
73%
51%
6%
99%
98%
94%
87%
76%
56%
8%
99%
98%
95%
89%
78%
60%
10%
99%
98%
95%
90%
80%
63%
12%
99%
98%
96%
90%
81%
66%
14%
99%
98%
96%
91%
83%
69%
16%
100%
99%
96%
92%
84%
72%
18%
100%
99%
97%
93%
86%
74%
20%
100%
99%
97%
93%
87%
76%
176
MARKETING METRICS
5.4 Prospect Lifetime Value Versus Customer Value
Prospect lifetime value is the expected value of a prospect. It is the value expected
from the prospect minus the cost of prospecting. The value expected from the prospect is the expected fraction of prospects who will make a purchase times the sum of
the average margin the firm makes on the initial purchase and the CLV of the newly
acquired customer.
Only if prospect lifetime value is positive should the firm proceed with the
planned acquisition spending.
Purpose: To account for the lifetime value of a newly acquired customer
(CLV) when making prospecting decisions.
One of the major uses of CLV is to inform prospecting decisions. A prospect is someone whom the firm will spend money on in an attempt to acquire her or him as a
customer. The acquisition spending must be compared not just to the contribution
from the immediate sales it generates but also to the future cash flows expected from
the newly acquired customer relationship (the CLV). Only with a full accounting of
the value of the newly acquired customer relationship will the firm be able to make an
informed, economic prospecting decision.
Construction
The expected prospect lifetime value (PLV) is the value expected from each prospect
minus the cost of prospecting. The value expected from each prospect is the acquisition
rate (the expected fraction of prospects who will make a purchase and become customers) times the sum of the initial margin the firm makes on the initial purchases and the
CLV. The cost is the amount of acquisition spending per prospect. The formula for
expected PLV is as follows:
Prospect Lifetime Value ($) = Acquisition Rate (%) * [Initial Margin ($) + CLV ($)]
– Acquisition Spending ($)
If PLV is positive, the acquisition spending is a wise investment. If PLV is negative, the
acquisition spending should not be made.
The PLV number will usually be very small. Although CLV is sometimes in the hundreds of dollars, PLV can come out to be only a few pennies. Just remember that PLV
applies to prospects, not customers. A large number of small but positive-value prospects can add to a considerable amount of value for a firm.
Chapter 5 Customer Profitability
177
EXAMPLE: A service company plans to spend $60,000 on an advertisement reaching 75,000
readers. If the service company expects the advertisement to convince 1.2% of the readers to take
advantage of a special introductory offer (priced so low that the firm makes only $10 margin on this
initial purchase) and the CLV of the acquired customers is $100, is the advertisement economically
attractive?
Here Acquisition Spending is $0.80 per prospect, the expected acquisition rate is 0.012, and the
initial margin is $10. The expected PLV of each of the 75,000 prospects is
PLV = 0.012 * ($10 + $100) – $0.80
= $0.52
The expected PLV is $0.52. The total expected value of the prospecting effort will be 75,000 * $0.52
= $39,000. The proposed acquisition spending is economically attractive.
If we are uncertain about the 0.012 acquisition rate, we might ask what the acquisition rate from the
prospecting campaign must be in order for it to be economically successful. We can get that number using Excel’s goal seek function to find the acquisition rate that sets PLV to zero. Or we can use
a little algebra and substitute $0 in for PLV and solve for the break-even acquisition rate:
Break-Even Acquisition Rate =
=
Acquisition Spending ($)
Initial Margin ($) + CLV ($)
$0.80
= 0.007273
$10 + $100
The acquisition rate must exceed 0.7273% in order for the campaign to be successful.
Data Sources, Complications, and Cautions
In addition to the CLV of the newly acquired customers, the firm needs to know
the planned amount of acquisition spending (expressed on a per-prospect basis), the
expected success rate (the fraction of prospects expected to become customers), and the
average margin the firm will receive from the initial purchases of the newly acquired
customers. The initial margin number is needed because CLV as defined in the previous section accounts for only the future cash flows from the relationship. The initial
cash flow is not included in CLV and must be accounted for separately. Note also that
the initial margin must account for any first-period retention spending.
Perhaps the biggest challenge in calculating PLV is estimating CLV. The other terms
(acquisition spending, acquisition rate, and initial margin) all refer to flows or outcomes in the near future, whereas CLV requires longer-term projections.
178
MARKETING METRICS
Another caution worth mentioning is that the decision to spend money on customer
acquisition whenever PLV is positive rests on an assumption that the customers
acquired would not have been acquired had the firm not spent the money. In other
words, our approach gives the acquisition spending “full credit” for the subsequent customers acquired. If the firm has several simultaneous acquisition efforts, dropping one
of them might lead to increased acquisition rates for the others. Situations such as these
(where one solicitation cannibalizes another) require a more complicated analysis.
The firm must be careful to search for the most economical way to acquire new customers. If there are alternative prospecting approaches, the firm must be careful not to
simply go with the first one that gives a positive projected PLV. Given a limited number
of prospects, the approach that gives the highest expected PLV should be used.
Finally, we want to warn you that there are other ways to do the calculations necessary to judge the economic viability of a given prospecting effort. Although these other
approaches are equivalent to the one presented here, they differ with respect to what
gets included in “CLV.” Some will include the initial margin as part of “CLV.” Others
will include both the initial margin and the expected acquisition cost per acquired customer as part of “CLV.” We illustrate these two approaches using the service company
example.
EXAMPLE: A service company plans to spend $60,000 on an advertisement reaching 75,000
readers. If the service company expects the advertisement to convince 1.2% of the readers to take
advantage of a special introductory offer (priced so low that the firm makes only $10 margin on this
initial purchase) and the CLV of the acquired customers is $100, is the advertisement economically
attractive?
If we include the initial margin in “CLV” we get
“CLV” [with Initial Margin ($)] = Initial Margin ($) + CLV ($)
= $10 + $110 = $110
The expected PLV is now
PLV ($) = Acquisition Rate (%) * “CLV” [with Initial Margin ($)] – Acquisition Cost ($)
= 0.012 * $110 – $0.85 = $0.52
This is the same number as before calculated using a slightly different “CLV”—one that includes the
initial margin.
We illustrate one final way to do the calculations necessary to judge the economics of a prospecting campaign. This last way does things on a per-acquired-customer basis using a “CLV” that
includes both initial margin and an allocated acquisition spending. The thinking goes as follows:
The expected value of a new customer is $10 now plus $100 from future sales, or $110 in total.
The expected cost to acquire a customer is the total cost of the campaign divided by the expected
Chapter 5 Customer Profitability
179
number of new customers. This average acquisition cost is calculated as $60,000/(0.012 * 75,000)
= $66.67. The expected value of a new customer net of the expected acquisition cost per customer
is $110 – $66.67 = $43.33. Because this new “net” CLV is positive, the campaign is economically
attractive. Some will even label this $43.33 number as the “CLV” of a new customer, although we
find this can create confusion.
Notice that $43.33 times the 900 expected new customers equals $39,000, the same total net value
from the campaign calculated in the original example as the $0.52 PLV times the 75,000 prospects.
The two ways to do the calculations are equivalent.
5.5 Acquisition Versus Retention Cost
The firm’s average acquisition cost is the ratio of acquisition spending to the number
of customers acquired. The average retention cost is the ratio of retention spending
directed toward a group of customers to the number of those customers successfully
retained.
Average Acquisition Cost ($) =
Average Retention Cost ($) =
Acquisition Spending ($)
Number of Customers Acquired (#)
Retention Spending ($)
Number of Customers Retained (#)
These two metrics help the firm monitor the effectiveness of two important categories of marketing spending.
Purpose: To determine the firm’s cost of acquisition and retention.
Before the firm can optimize its mix of acquisition and retention spending, it must first
assess the status quo. At the current spending levels, how much does it cost the firm (on
average) to acquire new customers, and how much is it spending (on average) to retain
its existing customers? Does it cost five times as much to acquire a new customer as it
does to retain an existing one?
Construction
Average Acquisition Cost: This represents the average cost to acquire a
customer and is the total acquisition spending divided by the number of new
customers acquired.
180
MARKETING METRICS
Average Acquisition Cost ($) =
Acquisition Spending ($)
Number of Customers Acquired (#)
Average Retention Cost: This represents the average “cost” to retain an existing
customer and is the total retention spending divided by the number of customers
retained.
Average Retention Cost ($) =
Retention Spending ($)
Number of Customers Retained (#)
EXAMPLE: During the past year, a regional pest control service spent $1.4 million and
acquired 64,800 new customers. Of the 154,890 customer relationships in existence at the start of
the year, only 87,957 remained at the end of the year, despite about $500,000 spent during the year
in attempts to retain the 154,890 customers. The calculation of average acquisition cost is relatively
straightforward. A total acquisition spend of $1.4 million resulted in 64,800 new customers. The
average acquisition cost is $1,400/64.8 = $21.60 per customer. The calculation of average retention
cost is also straightforward. A total retention spend of $500,000 resulted in 87,957 retained customers. The average yearly retention cost is $500,000/87,957 = $5.68. Thus, for the pest control firm, it
cost about four times as much to acquire a new customer as it did to retain an existing one.
Data Sources, Complications, and Cautions
For any specific period, the firm needs to know the total amount it spent on customer
acquisition and the number of new customers that resulted from that spending. With
respect to customer retention, the firm needs to measure the total amount spent during the period attempting to retain the customers in existence at the start of the period
and the number of the existing customers successfully retained at the end of the period.
Notice that retention spending directed at customers acquired within the period is not
included in this figure. Similarly, the number retained refers only to those retained from
the pool of customers in existence at the start of the period. Thus, the average retention
cost calculated will be associated with the length of the period in question. If the period
is a year, the average retention cost will be a cost per year per customer retained.
The calculation and interpretation of average acquisition cost is much easier than the
calculation and interpretation of average retention cost. This is so because it is often
possible to isolate acquisition spending and count the number of new customers
that resulted from that spending. A simple division results in the average cost to acquire
a customer. The reasonable assumption underlying this calculation is that the new
customers would not have been acquired had it not been for the acquisition spending.
Chapter 5 Customer Profitability
181
Things are not nearly so clear when it comes to average retention cost. One source of
difficulty is that retention rates (and costs) depend on the period of time under consideration. Yearly retention is different from monthly retention. The cost to retain a
customer for a month will be less than the cost to retain a customer for a year. Thus,
the definition of average retention cost requires a specification of the time period
associated with the retention.
A second source of difficulty stems from the fact that some customers will be retained
even if the firm spends nothing on retention. For this reason it can be a little misleading
to call the ratio of retention spending to the number of retained customers the average retention cost. One must not jump to the conclusion that retention goes away if
the retention spending goes away. Nor should one assume that if the firm increases the
retention budget by the average retention cost that it will retain one more customer.
The average retention cost number is not very useful to help make retention budgeting
decisions.
One final caution involves the firm’s capability to separate spending into acquisition
and retention classifications. Clearly there can be spending that works to improve both
the acquisition and retention efforts of the firm. General brand advertisements, for
example, serve to lower the cost of both acquisition and retention. Rather than attempt
to allocate all spending as either acquisition or retention, we suggest that it is perfectly
acceptable to maintain a separate category that is neither acquisition nor retention.
References and Suggested Further Reading
Berger, P.D., B. Weinberg, and R.C. Hanna. (2003). “Customer Lifetime Value Determination
and Strategic Implications for a Cruise-Ship Line,” Database Marketing and Customer Strategy
Management, 11(1), 40-52.
Blattberg, R.C., and S.J. Hoch. (1990). “Database Models and Managerial Intuition: 50% Model +
50% Manager,” Management Science, 36(8), 887–899.
Gupta, S., and Donald R. Lehmann. (2003). “Customers As Assets,” Journal of Interactive
Marketing, 17(1).
Kaplan, R.S., and V.G. Narayanan. (2001). “Measuring and Managing Customer Profitability,”
Journal of Cost Management, September/October: 5–15.
Little, J.D.C. (1970). “Models and Managers: The Concept of a Decision Calculus,” Management
Science, 16(8), B-466; B-485.
McGovern, G.J., D. Court, J.A. Quelch, and B. Crawford. (2004). “Bringing Customers into the
Boardroom,” Harvard Business Review, 82(11), 70–80.
182
MARKETING METRICS
Much, J.G., Lee S. Sproull, and Michal Tamuz. (1989). “Learning from Samples of One or
Fewer,” Organization Science: A Journal of the Institute of Management Sciences, 2(1), 1–12.
Peppers, D., and M. Rogers. (1997). Enterprise One-to-One: Tools for Competing in the Interactive
Age (1st ed.), New York: Currency Doubleday.
Pfeifer, P.E., M.E. Haskins, and R.M. Conroy. (2005). “Customer Lifetime Value, Customer
Profitability, and the Treatment of Acquisition Spending,” Journal of Managerial Issues, 17(1),
11–25.
Chapter 5 Customer Profitability
183
This page intentionally left blank
6
SALES FORCE AND CHANNEL
MANAGEMENT
Introduction
Key concepts covered in this chapter:
Sales Force Coverage
Facings and Share of Shelf
Sales Force Goals
Out-of-Stock and Service Levels
Sales Force Results
Inventory Turns
Sales Force Compensation
Markdowns
Pipeline Analysis
Gross Margin Return on Inventory
Investment (GMROII)
Numeric Distribution, ACV
Distribution, and PCV Distribution
Direct Product Profitability (DPP)
This chapter deals with push marketing. It describes how marketers measure the
adequacy and effectiveness of the systems that provide customers with opportunities to
buy their products.
The first sections discuss sales force metrics. Here, we list and define the most common measures for determining whether sales force effort and geographic coverage are
adequate. We discuss pipeline analysis, which is useful in making sales forecasts and
in allocating sales force effort to different stages of the selling process. Pipeline metrics are used to examine a sequence of selling activities, from lead generation, through
follow-up, to conversion and sales. Although the most important of these represents the
185
percentage of initial leads who ultimately buy, other measures of activity, productivity,
efficiency, and cost can be useful at each stage of the selling process.
In further sections of this chapter, we discuss measures of product distribution
and availability. For manufacturers who approach their market through resellers,
three key metrics provide an indication of “listings”—the percentage of potential outlets that stock their products. These include numeric distribution, which is
unweighted; ACV, the industry standard; and PCV, a category-specific measure of
product availability.
Marketing logistics tracking metrics are used to measure the operational effectiveness
of the systems that service retailers and distributors. Inventory turns, out-of-stocks, and
service levels are key factors in this area.
At the retail level, gross margin return on inventory investment (GMROII) and direct
product profitability (DPP) offer SKU-specific metrics of product performance, combining movement rates, gross margins, costs of inventory, and other factors.
Metric
Construction
Considerations
Purpose
6.1
Workload
Hours required to
service clients and
prospects.
Prospect numbers
may be debatable.
Time spent
trying to convert
prospects can
vary by territory,
salesperson,
and potential
client.
To assess the
number of
salespeople
required to service
a territory, and to
ensure balanced
workloads.
6.1
Sales Potential
Forecast
This comprises
the number of
prospects and their
buying power.
Doesn’t assess
the likelihood
of converting
“potential”
accounts.
Definitions of
buying power are
more an art than a
science.
To determine
sales targets. Can
also help identify
territories worthy
of an allocation
of limited sales
resources.
186
MARKETING METRICS
Metric
Construction
Considerations
Purpose
6.2
Sales Goal
Individual sales
projections may
be based on a
salesperson’s share
of forecasted sales,
on prior year sales
and a share of
increased district
projections, or on
a managementdesigned weighting
system.
Setting individual
targets on the
basis of prior
year sales can
discourage optimal
performance, as
strong performance
in one year leads
to more aggressive
targets in the next.
To set targets
for individual
salespeople and for
territories.
6.3
Sales Force
Effectiveness
Effectiveness
metrics analyze
sales in the context
of various criteria,
including calls,
contacts, potential
accounts, active
accounts, buying
power of territory,
and expenses.
Depends on factors
that also affect
sales potential and
workload.
To assess the
performance of
a salesperson
or team.
6.4
Compensation
Total payments
made to a
salesperson,
typically consisting
of base salary,
bonus, and/or
commission.
Perceived
relationship
between incentive
reward and
controllable
activities may vary
widely among
industries and
firms.
To motivate
maximum sales
effort. To enable
salespeople and
management to
track progress
toward goals.
6.4
Break-Even
Number of
Employees
Sales revenue,
multiplied by
margin net of
commission,
divided by cost per
staff member.
Margins may
vary across
products, time,
and salespeople.
Sales are not
independent of
the number of
salespeople.
To determine
the appropriate
personnel level for
a projected sales
volume.
Continues
Chapter 6 Sales Force and Channel Management
187
Metric
Construction
Considerations
6.5
Sales Funnel, Sales
Pipeline
Portrayal of the
number of clients
and potential
clients at various
stages of the sales
cycle.
Funnel dimensions To monitor sales
depend on type
effort and project
of business and
future sales.
definition of
potential clients.
6.6
Numeric
Distribution
Percentage of
outlets in a defined
universe that stock
a particular brand
or product.
Outlets’ size or
sales levels are
not reflected in
this measure.
Boundaries by
which distribution
universe is defined
may be arbitrary.
To assess the
degree to which a
brand or product
has penetrated its
potential channels.
6.6
All Commodity
Volume (ACV)
Numeric
distribution,
weighted by
penetrated outlets’
share of sales of all
product categories.
Reflects sales of
“all commodities,”
but may not
reflect sales of the
relevant product or
category.
To assess the
degree to which a
brand or product
has access to retail
traffic.
6.6
Product Category
Volume (PCV)
Numeric
distribution,
weighted
by penetrated
outlets’ share of
sales of the
relevant product
category.
Strong indicator
of share potential,
but may miss
opportunities to
expand category.
To assess the
degree to which a
brand or product
has access to
established outlets
for its category.
6.6
Total Distribution
Usually based on
ACV or PCV.
Sums the relevant
measures for
each SKU in a
brand or product
line.
Strong indicator
of the distribution
of a product line,
as opposed to an
individual SKU.
To assess the extent
to which a product
line is available.
188
MARKETING METRICS
Purpose
Metric
Construction
Considerations
Purpose
6.6
Category
The ratio of a PCV Same as for ACV
Performance Ratio to ACV
and PCV.
distribution.
To assess
whether a brand’s
distribution or a
particular retailer is
performing above
or below average
for the category.
6.7
Out-of-Stock
Percentage of
outlets that “list”
or normally stock a
product or brand,
but have none
available for sale.
Out-of-stocks can
be measured in
Numeric, ACV, or
PCV terms.
To monitor the
ability of logistics
systems to match
supply with
demand.
6.7
Inventories
Total amount of
product or brand
available for sale in
a channel.
May be held at
different levels and
valued in ways that
may or may not
reflect promotional
allowances and
discounts.
To calculate
ability to meet
demand and
determine channel
investments.
6.8
Markdowns
Percentage
discount from
the regular selling
price.
For many
products, a certain
percentage of
markdowns are
expected. Too
few markdowns
may reflect “underordering.” If
markdowns are too
high, the opposite
may be true.
To determine
whether channel
sales are being
made at planned
margins.
Continues
Chapter 6 Sales Force and Channel Management
189
Metric
Construction
Considerations
Purpose
6.8
Direct Product
The adjusted gross
Profitability (DPP) margin of products,
less direct product
costs.
Cost allocation is
often imprecise.
Some products may
be intended not to
generate profit but
to drive traffic.
To identify
profitable SKUs
and realistically
calculate their
earnings.
6.8
Gross Margin
Return on
Inventory
Investment
(GMROII)
Allowances and
rebates must
be considered
in margin
calculations. For
“loss leaders”
this measure may
be consistently
negative and still
not present a
problem. For most
products, negative
trends in GMROII
are signs of future
problems.
To quantify
return on working
capital invested in
inventory.
Margin divided by
the average dollar
value of inventory
held during a
specific period of
time.
6.1 Sales Force Coverage: Territories
Sales force territories are the customer groups or geographic districts for which
individual salespeople or sales teams hold responsibility. Territories can be defined
on the basis of geography, sales potential, history, or a combination of factors.
Companies strive to balance their territories because this can reduce costs and
increase sales.
Workload (#) = [Current Accounts (#) * Average Time to Service an Active Account (#)]
+ [Prospects (#) * Time Spent Trying to Convert a Prospect into
an Active Account (#)]
Sales Potential ($) = Number of Possible Accounts (#) * Buying Power ($)
190
MARKETING METRICS
Purpose: To create balanced sales territories.
There are a number of ways to analyze territories.1 Most commonly, territories are
compared on the basis of their potential or size. This is an important exercise. If territories differ sharply or slip out of balance, sales personnel may be given too much or
too little work. This can lead to under- or over-servicing of customers.
When sales personnel are stretched too thin, the result can be an under-servicing of
customers. This can cost a firm business because over-taxed salespeople engage in suboptimal levels of activity in a number of areas. They seek out too few leads, identify too
few prospects, and spend too little time with current customers. Those customers, in
turn, may take their business to alternate providers.
Over-servicing, by contrast, may raise costs and prices and therefore indirectly reduce
sales. Over-servicing in some territories may also lead to under-servicing in others.
Unbalanced territories also raise the problem of unfair distribution of sales potential
among members of a sales force. This may result in distorted compensation and cause
talented salespeople to leave a company, seeking superior balance and compensation.
Achieving an appropriate balance among territories is an important factor in maintaining satisfaction among customers, salespeople, and the company as a whole.
Construction
In defining or redefining territories, companies strive to
■
Balance workloads
■
Balance sales potential
■
Develop compact territories
■
Minimize disruptions during the redesign
These goals can have different effects on different stakeholders, as represented in
Table 6.1.2
Before designing new territories, a sales force manager should evaluate the workloads of all members of the sales team. The workload for a territory can be calculated
as follows:
Workload (#) = [Current Accounts (#) * Average Time to Service an Active Account (#)]
+ [Prospects (#) * Time Spent Trying to Convert a Prospect into an
Active Account (#)]
The sales potential in a territory can be determined as follows:
Sales Potential ($) = Number of Possible Accounts (#) * Buying Power ($)
Chapter 6 Sales Force and Channel Management
191
Table 6.1 Effects of Balancing Sales Territories
Balance the
Workload
Customers
Responsiveness
Relationships
Salespeople Earnings
opportunities
Manageable
workload
Reduced
uncertainty
Control of
overnights
Firm
Sales results
Effort control
Motivation
Travel cost
control
Balance
Sales
Potential
Minimize
Disruption
X
Develop
Compact
Territories
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Buying power is a dollar figure based on such factors as average income levels, number
of businesses in a territory, average sales of those businesses, and population demographics. Buying power indices are generally specific to individual industries.
EXAMPLE: Among the sales prospects in one of its territories, a copier manufacturer has
identified six small businesses, eight medium-sized firms, and two large companies. Enterprises
of these sizes have historically made annual copier purchases that average $500, $700, and $1,000,
respectively. The sales potential for the territory is thus:
Sales Potential = (6 * $500) + (8 * $700) + (2 * $1,000) = $10,600
In addition to workload and sales potential, a third key metric is needed to compare
territories. This is size or, more specifically, travel time. In this context, travel time is
more useful than size because it more accurately represents the factor that size implies—
that is, the amount of time needed to reach customers and potential customers.
As a manager’s goal is to balance workload and potential among sales personnel, it can
be beneficial to calculate combined metrics—such as sales potential or travel time—in
order to make comparisons between territories.
192
MARKETING METRICS
Data Sources, Complications, and Cautions
Sales potential can be represented in a number of ways. Of these, the most basic is
population—the number of potential accounts in a territory. In the copier case cited
earlier, this might be the number of offices in a territory.
Estimating the size of a territory might involve simply calculating the geographic area
that it covers. It is likely, however, that average travel time will also be important.
Depending on the quality of roads, density of traffic, or distance between businesses,
one may find that territories of equal area entail very different travel time requirements.
In evaluating such distinctions, sales force records of the time needed to travel from call
to call can be useful. Computer software programs, e.g., Google Maps, are available to
generate estimates.
Redefining territories is a famously difficult process. To perform it well, in addition to
the metrics cited earlier, disruption of customer relationships and feelings of ownership
among sales personnel must also be considered.
6.2 Sales Force Objectives: Setting Goals
Sales goals are generally needed to motivate salespeople. These can have negative
effects, however, if set too high or low. Means of establishing sales goals include the
following:
Sales Goal ($) = Salesperson’s Share of Prior-Year Sales in District (%)
* Forecasted Sales for District ($)
Sales Goal ($) = Salesperson’s Prior-Year Sales ($) + [Forecasted Sales Increase for
District ($) * Territory’s Share of Sales Potential in District (%)]
Weighted Share of Sales Allotment (%) = {Salesperson’s Share of Prior-Year Sales
in District (%) * Assigned Weighting (%)}
+ {Territory’s Share of Sales Potential in
District (%) * [1 – Assigned Weighting (%)]}
Sales Goal ($) = Weighted Share of Sales Allotment (%)
* Forecasted Sales for District ($)
Many of these approaches involve a combination of historical results and a weighting of sales potential among the territories. This ensures that overall goals will be
attained if all salespeople meet their individual goals.
Purpose: To motivate sales personnel and establish benchmarks
for evaluating and rewarding their performance.
In setting sales goals, managers strive to motivate their personnel to stretch themselves
and generate the most sales possible. But they don’t want to set the bar too high. The
correct goal levels will motivate all salespeople and reward most of them.
Chapter 6 Sales Force and Channel Management
193
When planning sales goals, certain guidelines are important. Under the SMART
strategy recommended by Jack D. Wilner, author of Seven Secrets to Successful Sales
Management,3 goals should be Specific, Measurable, Attainable, Realistic, and Timebound. Goals should be specific to a department, a territory, and even a salesperson.
They should be clear and applicable to each individual so that salespeople do not have
to derive part of their goal. Measurable goals, expressed in concrete numbers such as
“dollar sales” or “percentage increase,” enable salespeople to set precise targets and
track their progress. Vague goals, such as “more” or “increased” sales, are not effective
because they make it difficult to measure progress. Attainable goals are in the realm of
possibility. They can be visualized and understood by both the manager and the salesperson. Realistic goals are set high enough to motivate, but not so high that salespeople
give up before they even start. Finally, time-bound goals must be met within a precise
time frame. This applies pressure to reach them sooner rather than later and defines an
endpoint when results will be checked.
Construction
There are numerous ways of allotting a company’s forecast across its sales force. These
methods are designed to set goals that are fair, achievable, and in line with historic
results. Goals are stated in terms of sales totals for individual salespeople. In the following formulas, which encapsulate these methods, a district is composed of the individual
territories of multiple salespeople.
A sales goal or allocation based on prior-year sales can be calculated as follows:4
Sales Goal ($) =
Salesperson’s Share of Prior-Year Sales in District (%) * Forecasted Sales
for District ($)
A sales goal based on prior-year sales and the sales potential of a territory can be calculated as follows:
Sales Goal ($) = Salesperson’s Prior-Year Sales ($) + [Forecasted Sales Increase for District($)
* Territory’s Share of Sales Potential in District (%)]
Sales goals can also be set by a combined method, in which management assigns
weightings to both the prior-year sales of each salesperson and the sales potential of
each territory. These weightings are then used to calculate each salesperson’s percentage share of the relevant sales forecast, and percentage shares are used to calculate sales
goals in dollar terms.
Weighted Share of Sales Allotment (%) = {Salesperson’s Share of Prior-Year Sales in District (%)
* Assigned Weighting (%)} + {Territory’s Share
of Sales Potential in District (%) * [1 – Assigned
Weighting (%)]}
Sales Goal ($) = Weighted Share of Sales Allotment (%) * Forecasted Sales for District ($)
194
MARKETING METRICS
EXAMPLE: A salesperson achieved prior-year sales of $1,620, which represented 18% of the
sales in her district. This salesperson was responsible for a territory that held 12% of the sales potential in the district. If the salesperson’s employer mandates a district sales goal of $10,000 for the coming year—representing an overall increase of $1,000 over prior-year results—then the salesperson’s
individual sales goal can be calculated in several ways that involve different emphasis on historical
sales versus sales potential. Here are four examples:
1. Sales Goal Based on Prior-year Sales = 18% * $10,000 = $1,800
2. Sales Goals Based on Sales Potential = 12% * $10,000 = $1,200
3. Sales Goal Based on Prior-year Sales + Sales Potential * Increase = $1,620 + (12% * $1,000)
= $1,740
4. Weighted Share of Sales Allotment, in Which Prior-year Sales and Sales Potential Are
Weighted (for Example) by a Factor of 50% Each = (18% * 50%) + (12% * 50%) = 15%.
Then…
Sales Goal Based on Weighted Share of Sales Allotment = 15% * $10,000 = $1,500
Data Sources, Complications, and Cautions
Sales goals are generally established by using combinations of bottom-up and top-down
procedures. Frequently, top management sets objectives at a corporate level, while the
sales manager allocates shares of that overall goal among the various members of the sales
force.
Top management generally uses multiple metrics to forecast sales, including prior-year
sales of the product in question, total prior-year sales in the relevant market, prior-year
sales by competitors, and the company’s current market share. After the corporate sales
forecast is derived, a sales force manager verifies that these targets are reasonable, pushing back where necessary. The manager then allots the projected sales among the sales
force in a district, based at least in part on measures of individual performance from
the prior year. Of greatest importance in this calculation are each salesperson’s historic
percentage of sales and the sales potential of his or her territory.
It is important to re-evaluate sales goals during the year to ensure that actual performance
is running reasonably close to projections. If, at this checkpoint, it appears that more than
90% or less than 50% of the sales force is on track to achieve their goals, then it may be
advisable to alter the goals. This will prevent salespeople from easing off too early because
their goals are in sight, or giving up because their goals are unattainable. In setting goals,
one possible rule of thumb would be to plan for a success rate of 75%. That would ensure
that enough salespeople reach their goal and that the goal is sufficiently challenging.
If “rebudgeting” becomes necessary, it is important to ensure that this is properly
recorded. Unless care is taken, revised sales goals can slip out of alignment with financial budgets and the expectations of senior management.
Chapter 6 Sales Force and Channel Management
195
6.3 Sales Force Effectiveness: Measuring Effort,
Potential, and Results
By analyzing sales force performance, managers can make changes to optimize sales
going forward. Toward that end, there are many ways to gauge the performance of
individual salespeople and of the sales force as a whole, in addition to total annual
sales.
Sales Force Effectiveness Ratios
=
=
=
=
=
Sales ($)
Contacts with Clients (Calls) (#)
Sales ($)
Potential Accounts (#)
Sales ($)
Active Accounts (#)
Sales ($)
Buying Power ($)
Expenses ($)
(Also Known As Cost of Sales)
Sales ($)
Each can also be calculated on a dollar contribution basis.
Purpose: To measure the performance of a sales force
and of individual salespeople.
When analyzing the performance of a salesperson, a number of metrics can be compared. These can reveal more about the salesperson than can be gauged by his or her
total sales.
Construction
An authoritative source lists the following ratios as useful in assessing the relative effectiveness of sales personnel:5
Sales ($)
Contacts with Clients (Calls) (#)
Sales ($)
Potential Accounts (#)
196
MARKETING METRICS
Sales ($)
Active Accounts (#)
Sales ($)
Buying Power ($)
These formulas can be useful for comparing salespeople from different territories and
for examining trends over time. They can reveal distinctions that can be obscured by
total sales results, particularly in districts where territories vary in size, in number of
potential accounts, or in buying power.
These ratios provide insight into the factors behind sales performance. If an individual’s
sales per call ratio is low, for example, that may indicate that the salesperson in question needs training in moving customers toward larger purchases. Or it may indicate a
lack of closing skills. If the sales per potential account or sales per buying power metric
is low, the salesperson may not be doing enough to seek out new accounts. These metrics reveal much about prospecting and lead generation because they’re based on each
salesperson’s entire territory, including potential as well as current customers. The sales
per active account metric provides a useful indicator of a salesperson’s effectiveness in
maximizing the value of existing customers.
Although it is important to make the most of every call, a salesperson will not reach his
or her goal in just one call. A certain amount of effort is required to complete sales. This
can be represented graphically (see Figure 6.1).6
Although one can increase sales by expending more time and attention on a customer,
at a certain point, a salesperson encounters diminishing returns in placing more calls to
Sales ($)/Potential Account (#)
Calls (#)/Potential Account (#)
Figure 6.1
Sales Resulting from Calls to Customers
Chapter 6 Sales Force and Channel Management
197
the same customers. Eventually, the incremental business generated by each call will be
worth less than the cost of making the call.
In addition to the formulas described earlier, one other important measure of effectiveness is the ratio of expenses to sales. This cost metric is commonly expressed as a
percentage of sales and is calculated as follows:
Expenses ($)
Sales ($)
If this ratio is substantially higher for one salesperson than for others, it may indicate
that the individual in question has poor control of his or her expenses. Examples of
poor expense control could include making unnecessary trips to a client, overproducing
product pamphlets, or hosting too many dinners. Alternatively, expenses may represent
a high percentage of sales if an individual possesses poor closing skills. If a salesperson’s
expenses are comparable to those of his peers, but his sales are lower, then he may be
failing to deliver sales after spending significant money on a potential customer.
A more challenging set of sales force performance metrics involves customer service.
Customer service is difficult to measure because there are no concrete numbers representing it, other than repeat rates or customer complaints. Each of those is telling,
but how can a sales manager evaluate the service provided to customers who are not
repeating, leaving, or complaining? One possibility is to develop a survey, including an
itemized scale to help customers quantify their opinions. After enough of these surveys
are completed, managers will be able to calculate average scores for different service
metrics. By comparing these with sales figures, managers can correlate sales with customer service and grade salespeople on their performance.
EXAMPLE: To translate customers’ opinions into a metric, a company might pose survey
questions such as the following:
Please circle the level of service your business received from our sales staff after shipment of the
products you ordered:
1
2
3
Extremely Poor
4
5
6
7
Satisfactory
8
9
10
Extremely Good
Data Sources, Complications, and Cautions
Calculating the effectiveness of a salesperson is not difficult, but it does require keeping
track of a few important numbers. Fortunately, these are commonly tracked as part of
sophisticated sales management and customer relationship management (CRM) software packages.
198
MARKETING METRICS
The most important statistics are the amount of each sale (in dollars) and the contribution generated by that sale. It may also be important to keep track of which items are
sold if a salesperson has been instructed to emphasize a certain product line. Additional
useful information would include measures of the number of calls made (including
both face-to-face and phone meetings), total accounts active, and total accounts in the
territory. Of these, the latter two are needed to calculate the buying power of a territory.
The largest problem in performance review is a tendency to rely on only one or two
metrics. This can be dangerous because an individual’s performance on any one measure may be anomalous. A salesperson who generates $30,000 per call may be more
valuable than one who generates $50,000 per call, for example, if he generates greater
sales per potential account. A salesperson in a small territory may generate low total
contribution but high dollar sales per buying power. If this is true, it may be advisable
to increase the size of that person’s territory. Another salesperson may show a dramatic
increase in dollar sales per active account. If he achieves this simply by eliminating
weaker accounts without generating incremental sales, it would not be grounds for
reward. In reviewing sales personnel, managers are advised to evaluate as many performance metrics as possible.
Although the customer service survey described earlier is grounded upon a straightforward concept, managers can find it difficult to gather enough data—or sufficiently representative data—to make it useful. This could be because customers hesitate to fill out
the surveys, or because they do so only when they encounter a problem. A small sample
size or a prevalence of negative responses might distort the results. Even so, some effort
to measure customer satisfaction is needed to ensure that salespeople don’t emphasize
the wrong issues—or neglect issues that have a substantial impact on customers’ lifetime
value.
6.4 Sales Force Compensation: Salary/Reward Mix
“The incentive plan needs to align the salesperson’s activities with the firm’s objectives.”7 Toward that end, an effective plan may be based on the past (growth), the
present (comparison with others), or the future (percentage of goal achieved). Key
formulas in this area include the following:
Compensation ($) = Salary ($) + Bonus 1 ($) + Bonus 2 ($)
Compensation ($) = Salary ($) + [Sales ($) * Commission (%)]
Break-Even Number of Employees (#) =
(Sales ($) * [Margin (%) – Commission (%)])
[Salary ($) + Expenses ($) + Bonus ($)]
Chapter 6 Sales Force and Channel Management
199
Purpose: To determine the mix of salary, bonus, and commission that will
maximize sales generated by the sales force.
When designing a compensation plan for a sales force, managers face four key considerations: level of pay, mix between salary and incentive, measures of performance, and
performance-payout relationships. The level of pay, or compensation, is the amount
that a company plans to pay a salesperson over the course of a year. This can be viewed
as a range because its total will vary with bonuses or commissions.
The mix between salary and incentive represents a key allocation within total compensation. Salary is a guaranteed sum of money. Incentives can take multiple forms,
including bonuses or commissions. In the case of a bonus, a salesperson will receive
a lump sum for reaching certain sales targets. With a commission, the incentive is
incremental and is earned on each sale. In order to generate incentives, it is important
to measure accurately the role a salesperson plays in each sale. The higher the level
of causality that can be attributed to a salesperson, the easier it is to use an incentive
system.
Various metrics can be used to measure a salesperson’s performance. With these, managers can evaluate a salesperson’s performance in the context of past, present, or future
comparators, as follows:
■
■
■
The past: Measure the salesperson’s percentage growth in sales over prior-year
results.
The present: Rank salespeople on the basis of current results.
The future: Measure the percentage of individual sales goals achieved by each
salesperson.
Sales managers can also select the organizational level on which to focus an incentive
plan. The disbursement of incentive rewards can be linked to results at the company,
division, or product-line level. In measuring performance and designing compensation
plans along all these dimensions, managers seek to align salespeople’s incentives with
the goals of their firm.
Lastly, a time period should be defined for measuring the performance of each
salesperson.
Construction
Managers enjoy considerable freedom in designing compensation systems. The key is
to start with a forecast for sales and a range within which each salesperson’s compensation should reside. After these elements are determined, there are many ways to motivate a salesperson.
200
MARKETING METRICS
In a multi-bonus system, the following formula can represent the compensation structure for a salesperson:
Compensation ($) = Salary ($) + Bonus 1 ($) + Bonus 2 ($)
In this system, bonus 1 might be attained at a level approximately halfway to the individual’s sales goal for the year. The second bonus might be awarded when that goal is met.
In a commission system, the following formula would represent compensation for
a salesperson:
Compensation ($) = Salary ($) + [Sales ($) * Commission (%)]
Theoretically, in a 100% commission structure, salary might be set as low as $0. Many
jurisdictions, however, place limits on such arrangements. Managers must ensure that
their chosen compensation structures comply with employment law.
Managers can also combine bonus and commission structures by awarding bonuses
on top of commissions at certain sales levels, or by increasing the commission rate at
certain sales levels.
EXAMPLE: Tina earns a commission of 2% on sales up to $1,000,000, and a 3% commission
on sales beyond that point. Her salary is $20,000 per year. If she makes $1,200,000 in sales, her compensation can be calculated as follows:
Compensation = $20,000 + (.02) * ($1,000,000) + (.03) * ($200,000)
= $46,000
After a sales compensation plan has been established, management may want to reevaluate the size of its sales force. Based on forecasts for the coming year, a firm may
have room to hire more salespeople, or it may need to reduce the size of the sales force.
On the basis of a given value for projected sales, managers can determine the breakeven number of employees for a firm as follows:
Break-Even Number of Employees (#) =
Sales ($) * [Margin (%) – Commission (%)]
[Salary ($) + Expenses ($) + Bonus ($)]
Data Sources, Complications, and Cautions
Measurements commonly used in incentive plans include total sales, total contribution, market share, customer retention, and customer complaints. Because such a plan
rewards a salesperson for reaching certain goals, these targets must be defined at the
Chapter 6 Sales Force and Channel Management
201
beginning of the year (or other time period). Continual tracking of these metrics will
help both the salesperson and the company to plan for year-end compensation.
Timing is an important issue in incentive plans. A firm must collect data in a timely
fashion so that both managers and salespeople know where they stand in relation
to established goals. The time frame covered by a plan also represents an important
consideration. If a company tries to generate incentives through weekly rewards, its
compensation program can become too expensive and time-consuming to maintain.
By contrast, if the program covers too long a period, it may slip out of alignment with
company forecasts and goals. This could result in a sales force being paid too much or
too little. To guard against these pitfalls, managers can develop a program that mixes
both short- and long-term incentives. They can link some rewards to a simple, shortterm metric, such as calls per week, and others to a more complex, long-term target,
such as market share achieved in a year.
A further complication that can arise in incentive programs is the assignment of causality to individual salespeople. This can become a problem in a number of instances,
including team collaborations in landing sales. In such a scenario, it can be difficult to
determine which team members deserve which rewards. Consequently, managers may
find it best to reward all members of the team with equal bonuses for meeting a goal.
A last concern: When an incentive program is implemented, it may reward the “wrong”
salespeople. To avoid this, before activating any newly proposed program, sales managers are advised to apply that program to the prior year’s results as a test. A “good” plan
will usually reward the salespeople whom the manager knows to be the best.
6.5 Sales Force Tracking: Pipeline Analysis
Pipeline analysis is used to track the progress of sales efforts in relation to all current
and potential customers in order to forecast short-term sales and to evaluate sales
force workload.
Purpose: To forecast upcoming sales and evaluate workload distribution.
A convenient way to forecast sales in the short term and to keep an eye on sales force
activity is to create a sales pipeline or sales funnel. Although this concept can be
represented graphically, the data behind it are stored electronically in a database or
spreadsheet.
The concept of the sales funnel originates in a well-known dynamic: If a sales force
approaches a large number of potential customers, only a subset of these will actually make purchases. As salespeople proceed through multiple stages of customer
202
MARKETING METRICS
interaction, a number of prospects are winnowed out. At the conclusion of each stage,
fewer potential customers remain. By keeping track of the number of potential customers at each stage of the process, a sales force manager can balance the workload within
a team and make accurate forecasts of sales.
This analysis is similar to the hierarchy of effects discussed in Section 2.7. Whereas the
hierarchy of effects focuses on the impact of advertising or mass media, the sales funnel is
used to track individual customers (often by name) and sales force efforts. (Note: In some
industries, such as consumer packaged goods, the term “pipeline sales” can refer to sales
into a distribution channel. Please do not confuse pipeline sales with a sales pipeline.)
Construction
In order to conceptualize a sales funnel or pipeline, it is helpful to draw a diagram
showing the stages of the selling process (see Figure 6.2). At any point in the year, it is
likely that all stages of the pipeline will include some number of customers. As Figure
6.2 illustrates, although there may be a large number of potential customers, those who
actually make purchases represent only a percentage of these original leads.
冧
Cold Leads
Warm Leads
Prospects
Interest
Creation
冧
Pre-purchase
冧
Purchase
冧
Post-purchase
1st Meeting
2nd Meeting
3rd Meeting
Delivery
Support
Figure 6.2
Sales Force Funnel
Chapter 6 Sales Force and Channel Management
203
Interest Creation: This entails building awareness of a product through such activities
as trade shows, direct mail, and advertising. In the course of interest creation, salespeople can also generate leads. That is, they can identify targets to add to their pool
of potential customers. Two main classifications of leads include cold leads and warm
leads.
Cold Lead: A lead that has not specifically expressed interest. These can be identified through mailing lists, phone books, business listings, and so on.
Warm Lead: A lead that is expected to be responsive. These potential customers
may have registered through a Web site or requested product information, for
example.
Pre-Purchase: This stage involves identifying prospects from among cold and warm
leads. Salespeople make this distinction through initial meetings with leads, in which
they explain product features and benefits, and cooperate in problem solving with the
customer. The desired result of such an early-stage meeting is not a sale but rather the
identification of a prospect and the scheduling of another meeting.
Prospect: A potential customer who has been identified as a likely buyer, possessing the ability and willingness to buy.8
Purchase: After prospects are identified and agree to additional calls, salespeople
engage in second and third meetings with them. It is in these sessions that traditional
“selling” takes place. Salespeople will engage in persuading, negotiating, and/or bidding. If a purchase is agreed upon, a salesperson can close the deal through a written
proposal, contract, or order.
Post-Purchase: After a customer has made a purchase, there is still considerable work
to be done. This includes delivery of the product or service, installation (if necessary),
collection of payments, and possibly training. There is then an ongoing commitment to
customer service.
After salespeople visualize the different stages represented in a sales funnel, they can
track their customers and accounts more accurately. They can do this electronically by
using a database or spreadsheet. If a sales pipeline file is maintained on a shared drive,
any member of a sales force will be able to update the relevant data on a regular basis.
This will also enable a sales manager to view the progress of the team at any point in
time. Table 6.2 is an example of a spreadsheet form of a sales funnel.
A manager can use the information stored in such a funnel to prepare for sales in the
near future. This is a form of pipeline analysis. When a firm faces inventory issues,
or when sales goals are being missed, this represents vital information. By applying
historical averages, a sales or marketing manager can improve sales forecasts by using
the data in a sales funnel. This can be done manually or with specialized software. The
204
MARKETING METRICS
Table 6.2
Interest Creation
Spreadsheet Sales Funnel
Pre-purchase
1st/2nd
Meeting
Purchase
2nd/3rd
Meeting
Post-purchase
Cold
Leads
Warm
Leads
Sandy
56
30
19
5
8
7
25
Bob
79
51
33
16
4
14
35
Salesperson
Prospects
Delivery
Support
underlying assumption behind a sales funnel is that failure at any stage eliminates a
prospect from the funnel. The following example illustrates how this bottom-up forecasting could be applied.
EXAMPLE: Using the sales funnel from earlier, Sandy and Bob’s manager wants to forecast the
number of sales that will require fulfillment in the next five months. Toward that end, she applies
certain historical averages:
■
■
■
■
■
2% of cold calls are converted to sales within five months.
14% of warm calls are converted to sales within four months.
25% of prospects are converted to sales within three months.
36% of customers who agree to a pre-purchase meeting are converted to sales within
two months.
53% of customers who agree to a purchase meeting are converted to sales within
one month.
On this basis:
Upcoming Sales = [(56 + 79) * 2%] + [(30 + 51) * 14%] + [(19 + 33) * 25%]
+ [(5 + 16) * 36%)] + [(8 + 4) * 53%] = 41
Note: This example applies to only one product. Often, a firm will need multiple sales funnels for
different products or product lines. Additionally, a sale may comprise a single item or thousands
of items. In the latter case, it would be appropriate to use a metric for “average sale size/customer”
in forecasting.
Data Sources, Complications, and Cautions
In order to populate a sales funnel correctly, salespeople must maintain records of all
their current and potential customers, and the status of each within the purchase process. Each salesperson must also share this information, which can then be aggregated in
Chapter 6 Sales Force and Channel Management
205
a comprehensive database of sales force activities. By applying assumptions to these—
including assumptions drawn from historical sales results—a firm can project future
sales. For example, if 25% of warm leads are generally converted to sales within two
months, and 200 warm leads currently appear in a sales funnel, management can estimate that 50 of these will be converted to sales within two months.
At times, the use of a sales funnel leads to the pitfall of over-prospecting. If the incremental contribution generated by a customer is less than the cost of acquiring that
customer, then prospecting for that customer yields a negative result. Salespeople are
advised to use customer lifetime value metrics as a guide in deciding the appropriate
scale and direction of their prospecting. Increasing pre-purchase sales funnel metrics
will not be worthwhile unless that increment leads to improved figures further down
the pipeline as well.
Difficulties in the sales cycle can also arise when a salesperson judges that a potential
customer may be a prospect because he or she has the willingness and ability to buy.
To solidify this judgment, the salesperson must also confirm that the customer possesses
the authority to buy. When prospecting, salespeople should take the time needed to
verify that their contacts can make purchase decisions without approval from another
source.
6.6 Numeric, ACV and PCV Distribution,
Facings/Share of Shelf
Distribution metrics quantify the availability of products sold through resellers, usually as a percentage of all potential outlets. Often, outlets are weighted by their share
of category sales or “all commodity” sales.
Numeric Distribution (%) =
Number of Outlets Carrying Brand (#)
Total Number of Outlets (#)
All Commodity Volume (ACV) Distribution (%) =
Total Sales of Outlets Carrying Brand ($)
Product Category Volume (PCV) Distribution9 (%) =
Total Sales of All Outlets ($)
Total Category Sales of Outlets
Carrying Brand ($)
Category Performance Ratio (%) =
206
MARKETING METRICS
Total Category Sales of All
Outlets ($)
PCV (%)
ACV (%)
For marketers who sell through resellers, distribution metrics reveal a brand’s percentage of market access. Balancing a firm’s efforts in “push” (building and maintaining reseller and distributor support) and “pull” (generating customer demand) is
an ongoing strategic concern for marketers.
Purpose: To measure a firm’s ability to convey a product to its customers.
In broad terms, marketing can be divided into two key challenges:
■
■
The first—and most widely appreciated—is to ensure that consumers or end
users want a firm’s product. This is generally termed pull marketing.
The second challenge is less broadly recognized, but often just as important.
Push marketing ensures that customers are given opportunities to buy.
Marketers have developed numerous metrics by which to judge the effectiveness of the
distribution system that helps create opportunities to buy. The most fundamental of
these are measures of product availability.
Availability metrics are used to quantify the number of outlets reached by a product,
the fraction of the relevant market served by those outlets, and the percentage of total
sales volume in all categories held by the outlets that carry the product.
Construction
There are three popular measures of distribution coverage:
1. Numeric distribution
2. All commodity volume (ACV)
3. Product category volume (PCV), also known as weighted distribution
Numeric Distribution
This measure is based on the number of outlets that carry a product (that is, outlets
that list at least one of the product’s stock-keeping units, or SKUs). It is defined as the
percentage of stores that stock a given brand or SKU, within the universe of stores in
the relevant market.
The main use of numeric distribution is to understand how many physical locations
stock a product or brand. This has implications for delivery systems and for the cost of
servicing these outlets.
Chapter 6 Sales Force and Channel Management
207
Numeric Distribution: To calculate numeric distribution, marketers divide the number of stores that stock at least one SKU of a product or brand by the number of outlets
in the relevant market.
Numeric Distribution (%) =
Number of Outlets Carrying Product (#)
Total Number of Outlets in the Market (#)
For further information about stock-keeping units (SKUs), refer to Section 3.3.
EXAMPLE: Alice sells photo albums to gift shops. There are 60 such stores in her area. In order
to generate adequate distribution coverage, Alice believes she must reach at least 60% of these. In
initiating her relationship with each store, however, Alice must provide the store with $4,000 worth
of inventory to build a presence. To attain her distribution goal, how much will Alice need to invest
in inventory?
To reach her numeric distribution target of 60%, Alice must build a presence in 36 stores (that is,
0.60 * 60).
She will therefore have to spend at least $144,000 on inventory (36 stores * $4,000 per store).
All Commodity Volume
All commodity volume (ACV) is a weighted measure of product availability, or
distribution, based on total store sales. ACV can be expressed as a dollar value or
percentage.
All Commodity Volume (ACV): The dollar value of store sales in all categories
by stores in question, or the percentage in comparison with all stores in the
relevant universe.
All Commodity Volume (ACV Distribution) (%)
=
Total Sales of Stores Carrying Brand ($)
Total Sales of All Stores ($)
All Commodity Volume (ACV Distribution) ($) = Total Sales of Stores Carrying Brand ($)
EXAMPLE: The marketers at Madre’s Tortillas want to know the all commodity volume of
their distribution network (Table 6.3).
208
MARKETING METRICS
Table 6.3 Madre’s Tortillas’ Distribution
Outlet
All Sales
Tortilla Sales
Madre’s Tortillas
SKUs Stocked
Padre’s Tortillas
SKUs Stocked
Store 1
$100,000
$1,000
12 ct, 24 ct
12 ct, 24 ct
Store 2
$75,000
$500
12 ct
24 ct
Store 3
$50,000
$300
12 ct, 24 ct
none
Store 4
$40,000
$400
none
12 ct, 24 ct
Madre’s Tortillas are carried by Stores 1-3, but not by Store 4. The ACV of its distribution network
is therefore the total sales of Stores 1, 2, and 3, divided by the total sales of all stores. This represents
a measure of the sales of all commodities in these stores, not just tortilla sales.
Madre’s Tortillas ACV (%) =
=
=
Sales Stores 1 – 3
All Store Sales
($100k + $75k + $50k)
($100k + $75k + $50k + $40k)
$225k
$265k = 84.9%
The principal benefit of the ACV metric, by comparison with numeric distribution,
is that it provides a superior measure of customer traffic in the stores that stock a
brand. In essence, ACV adjusts numeric distribution for the fact that not all retailers generate the same level of sales. For example, in a market composed of two small
stores, one superstore, and one kiosk, numeric distribution would weight each outlet
equally, whereas ACV would place greater emphasis on the value of gaining distribution in the superstore. In calculating ACV when detailed sales data are not available,
marketers sometimes use the square footage of stores as an approximation of their total
sales volume.
The weakness of ACV is that it does not provide direct information about how
well each store merchandises and competes in the relevant product category. A store
can do a great deal of general business but sell very little of the product category under
consideration.
Chapter 6 Sales Force and Channel Management
209
Product Category Volume
Product category volume (PCV)10 is a refinement of ACV. In many countries, this metric is known simply as “weighted distribution.” It examines the share of the relevant
product category sold by the stores in which a given product has gained distribution. It
helps marketers understand whether a given product is gaining distribution in outlets
where customers look for its category, as opposed to simply high-traffic stores where
that product may get lost in the aisles.
Continuing our example of the two small retailers, the kiosk, and the superstore,
although ACV may lead the marketer of a chocolate bar to seek distribution in the hightraffic superstore, PCV might reveal that the kiosk, surprisingly, generates the greatest
volume in snack sales. In building distribution, the marketer would then be advised to
target the kiosk as her highest priority.
Product Category Volume (PCV): The percentage share, or dollar value, of category sales made by stores that stock at least one SKU of the brand in question, in
comparison with all stores in their universe.
Product Category Volume (PCV Distribution) (%) =
Total Category Sales by Stores
Carrying Brand ($)
Total Category Sales of All Stores ($)
Product Category Volume (PCV Distribution) ($) = Total Category Sales of Stores
Carrying Brand ($)
When detailed sales data are available, PCV can provide a strong indication of the
market share within a category to which a given brand has access. If sales data are not
available, marketers can calculate an approximate PCV by using square footage devoted
to the relevant category as an indication of the importance of that category to a particular outlet or store type.
EXAMPLE: The marketers at Madre’s Tortillas want to know how effectively their product is
reaching the outlets where customers shop for tortillas. Using data from the previous example:
Stores 1, 2, and 3 stock Madre’s Tortillas. Store 4 does not. The product category volume of Madre’s
Tortillas’ distribution network can be calculated by dividing total tortilla sales in Stores 1-3 by tortilla sales throughout the market.
PCV (%) =
=
210
Tortilla Sales of Stores Carrying Madre’s
Tortilla Sales of All Stores
$1,000 + $500 + $300
$1,000 + $500 + $300 + $400
MARKETING METRICS
= $81.8%
Total Distribution
The sum of ACV or PCV distribution for all of a brand’s stock-keeping-units, calculated individually. In many product categories, perhaps most, it is important to have
more than one stock-keeping-unit (SKU) available for purchase. Product variants can
include flavors, sizes, package types, formulations, or many other variations on the
basic product. Measuring ACV% or PCV% distribution for the brand is rarely sufficient
as an indicator of the product line’s availability. Total Distribution is the sum of the relevant distribution metric across all of a brand’s SKUs. For example, consider that Louis’
Ice Cocoa Energy Drink (LICED) is offered in four variations: 32oz chocolate, 32oz
vanilla, 16oz chocolate, and 16oz vanilla. The PCV% distribution for these four LICED
SKUs are: 80%, 70%, 60%, and 40% respectively. The Total Distribution for LICED is
the sum of the SKU distribution or 250%. If the LICED brand PCV distribution is 90%
it would mean that outlets stocking at least one SKU of LICED accounted for 90% of
energy drink sales. It would also mean that, on average, stocking outlets would have
about 2.8 SKUs on the shelf (250%/90%). Total distribution is a metric that combines
breadth of distribution for the brand and depth of distribution for the product line.
As brands have proliferated SKUs and variants it has become more important to track
Total Distribution and to compare the Total Distribution for the brand to the same
metric for the category. Consider, for example, the number of craft beers now available
in your local supermarket or liquor store.
Category Performance Ratio: The relative performance of a retailer or retailers in a given product category, compared with performance in all product
categories.
By comparing PCV with ACV, the category performance ratio provides insight into
whether a brand’s distribution network is more or less effective in selling the category
of which that brand is a part, compared with its average effectiveness in selling all categories in which members of that network compete.
Category Performance Ratio (%) =
PCV (%)
ACV (%)
If a distribution network’s category performance ratio is greater than 1, then the outlets
comprising that network perform comparatively better in selling the category in question than in selling other categories, relative to the market as a whole.
EXAMPLE: As noted earlier, the PCV of Madre’s Tortillas’ distribution network is 81.8%. Its
ACV is 84.9%. Thus, its category performance ratio is 0.96.
Madre’s has succeeded in gaining distribution in the largest stores in its market. Tortilla sales in
those stores, however, run slightly below the average of all commodity sales in those stores, relative
to the market as a whole. That is, outlets carrying Madre’s show a slightly weaker focus on tortillas
than the overall universe of stores in this market.
Chapter 6 Sales Force and Channel Management
211
Data Sources, Complications, and Cautions
In many markets, there are data suppliers such as A.C. Nielsen, which specialize in collecting information about distribution. In other markets, firms must generate their own
data. Sales force reports and shipment invoices provide a place to start.
For certain merchandise—especially low-volume, high-value items—it is relatively
simple to count the limited number of outlets that carry a given product. For highervolume, lower-cost goods, merely determining the number of outlets that stock an
item can be a challenge and may require assumptions. Take, for instance, the number of outlets selling a specific soft drink. To arrive at an accurate number, one
would have to include vending machines and street vendors as well as traditional
grocery stores.
Total outlet sales are often approximated by quantifying selling space (measured in
square feet or square meters) and applying this measure to industry averages for sales
per area of selling space.
In the absence of specific category sales data, it is often useful to weight ACV to arrive at
an approximation of PCV. Marketers may know, for example, that pharmacies, relative
to their overall sales, sell proportionally more of a given product than do superstores.
In this event, they might increase the weighting of pharmacies relative to superstores in
evaluating relevant distribution coverage.
Related Metrics and Concepts
Facing: A facing is a frontal view of a single package of a product on a fully stocked
shelf.
Share of Shelf: A metric that compares the facings of a given brand to the total facing
positions available in the category, in order to quantify the display prominence of that
brand.
Share of Shelf (%) =
Facings for Brand (#)
Total Facings in the Category(#)
Store Versus Brand Measures: Marketers often refer to a grocery chain’s ACV. This
can be either a dollar number (the chain’s total sales of all categories in the relevant
geographic market) or a percentage number (its share of dollar sales among the universe of stores). A brand’s ACV is simply the sum of the ACVs of the chains and stores
that stock that brand. Thus, if a brand is stocked by two chains in a market, and these
chains have 40% and 30% ACV respectively, then the ACV of that brand’s distribution
network is 30% + 40%, or 70%.
212
MARKETING METRICS
Marketers can also refer to a chain’s market share in a specific category. This is equivalent to the chain’s PCV (%). A brand’s PCV, by contrast, represents the sum of the
PCVs of the chains that stock that brand.
Inventory: This is the level of physical stock held. It will typically be measured at different points in a pipeline. A retailer may have inventory on order from suppliers, at
warehouses, in transit to stores, in the stores’ backrooms, and on the store shelves.
Breadth of Distribution: This figure can be measured by the number of SKUs held.
Typically, a company will hold a wide range of SKUs—a high breadth of distribution—
for the products that it is most interested in selling.
Features in Store: The percentage of stores offering a promotion in a given time period.
This can be weighted by product or by all commodity volume (ACV).
ACV on Display: Distinctions can be made in all commodity volume metrics to take
account of where products are on display. This will reduce the measured distribution of
products if they are not in a position to be sold.
ACV on Promotion: Marketers may want to measure the ACV of outlets where a given
product is on promotion. This is a useful shorthand way of determining the product’s
reliance on promotion.
6.7 Supply Chain Metrics
Marketing logistics tracking includes the following metrics:
Out-of-Stocks (%) =
Outlets Where Brand or Product Is Listed But Unavailable (#)
Total Outlets Where Brand or Product Is Listed (#)
Service Levels; Percentage on Time Delivery (%) =
Inventory Turns (I) =
Deliveries Achieved in
Timeframe Promised (#)
All Deliveries Initiated in the
Period (#)
Product Revenues ($)
Average Inventory ($)
Logistics tracking helps ensure that companies are meeting demand efficiently and
effectively.
Chapter 6 Sales Force and Channel Management
213
Purpose: To monitor the effectiveness of an organization in managing the
distribution and logistics process.
Logistics are where the marketing rubber meets the road. A lot can be lost at the potential point-of-purchase if the right goods are not delivered to the appropriate outlets on
time and in amounts that correspond to consumer demand. How hard can that be?
Well, ensuring that supply meets demand becomes more difficult when:
■
■
The company sells more than a few stock keeping units (SKUs).
Multiple levels of suppliers, warehouses, and stores are involved in the distribution process.
■
Product models change frequently.
■
The channel offers customer-friendly return policies.
In this complex field, by monitoring core metrics and comparing these with historical
norms and guidelines, marketers can determine how well their distribution channel is
functioning as a supply chain for their customers.
By monitoring logistics, managers can investigate questions such as the following: Did
we lose sales because the wrong items were shipped to a store that was running a promotion? Are we being forced to pay for the disposal of obsolete goods that stayed too
long in warehouses or stores?
Construction
Out-of-Stocks: This metric quantifies the number of retail outlets where an item
is expected to be available for customers, but is not. It is typically expressed as a
percentage of stores that list the relevant item.
Out-of-Stocks (%) =
Outlets Where Brand or Product Is Listed But Unavailable (#)
Total Outlets Where Brand or Product Is Listed (#)
Being “listed” by a chain means that a headquarters buyer has “authorized” distribution
of a brand, SKU, or product at the store level. For various reasons, being listed does not
always ensure presence on the shelf. Local managers may not approve “distribution.”
Alternatively, a product may be distributed but sold out.
Out-of-stocks are often expressed as a percentage. Marketers must note whether an
out-of-stock percentage is based on numeric distribution, ACV, PCV, or the percentage
of distributing stores for a given chain.
The in-stock percentage is the complement of the out-of-stock percentage. A 3% outof-stock rate would be equivalent to a 97% in-stock rate.
214
MARKETING METRICS
PCV Net Out-of-Stocks: The PCV of a given product’s distribution network,
adjusted for out-of-stock situations.
Product Category Volume (PCV), Net Out-of-Stocks: This out-of-stocks measure is
calculated by multiplying PCV by a factor that adjusts it to recognize out-of-stock situations. The adjusting factor is simply one minus the out-of-stocks figure.
Product Category Volume, Net Out-of-Stocks (%) = PCV (%) * [1 – Out-of-Stock (%)]
Service Levels, Percentage On-time Delivery: There are various service measures in
marketing logistics. One particularly common measure is on-time delivery. This metric
captures the percentage of customer (or trade) orders that are delivered in accordance
with the promised schedule.
Service Levels, Percentage on Time Delivery (%) =
Deliveries Achieved in Timeframe
Promised (#)
All Deliveries Initiated in the Period (#)
Inventories, like out-of-stocks and service levels, should be tracked at the SKU level.
For example, in monitoring inventory, an apparel retailer will need to know not only
the brand and design of goods carried, but also their size. Simply knowing that there
are 30 pairs of suede hiking boots in a store, for example, is not sufficient—particularly
if all those boots are the same size and fail to fit most customers.
By tracking inventory, marketers can determine the percentage of goods at each stage
of the logistical process—in the warehouse, in transit to stores, or on the retail floor, for
example. The significance of this information will depend on a firm’s resource management strategy. Some firms seek to hold the bulk of their inventory at the warehouse
level, for example, particularly if they have an effective transport system to ship goods
quickly to stores.
Inventory Turns: The number of times that inventory “turns over” in a year can be
calculated on the basis of the revenues associated with a product and the level of inventory held. One need only divide the revenues associated with the product in question
by the average level of inventory for that item. As this quotient rises, it indicates that
inventory of the item is moving more quickly through the process. Inventory turns can
be calculated for companies, brands, or SKUs and at any level in the distribution chain,
but they are frequently most relevant for individual trade customers. Important note: In
calculating inventory turns, dollar figures for both sales and inventory must be stated
either on a cost or wholesale basis, or on a retail or resale basis, but the two bases must
not be mixed.
Inventory Turns (I) =
Annual Product Revenues ($)
Average Inventory ($)
Chapter 6 Sales Force and Channel Management
215
Inventory Days: This metric also sheds light on the speed with which inventory moves
through the sales process. To calculate it, marketers divide the 365 days of the year by
the number of inventory turns, yielding the average number of days of inventory carried by a firm. By way of example, if a firm’s inventory of a product “turned” 36.5 times
in a year, that firm would, on average, hold 10 days’ worth of inventory of the product.
High inventory turns—and, by corollary, low inventory days—tend to increase profitability through efficient use of a firm’s investment in inventory. But they can also lead
to higher out-of-stocks and lost sales.
Inventory Days (#) =
Days in Year (365)
Inventory Turns (I)
Inventory days represents the number of days’ worth of sales that can be supplied by
the inventory present at a given moment. Viewed from a slightly different perspective,
this figure advises logistics managers of the time expected to elapse before they suffer a
stock-out. To calculate this figure, managers divide product revenue for the year by the
value of the inventory days, generating expected annual turns for that inventory level.
This can be easily converted into days by using the previous equation.
EXAMPLE: An apparel retailer holds $600,000 worth of socks in inventory January 1, and
$800,000 the following December 31. Revenues generated by sock sales totaled $3.5 million during
the year.
To estimate average sock inventory during the year, managers might take the average of the beginning and ending numbers: ($600,000 + $800,000)/2 = $700,000 average inventory. On this basis,
managers might calculate inventory turns as follows:
Inventory Turns =
=
Product Revenues
Average Inventory
$3,500,000
$700,000
=5
This figure can be converted to inventory days in order to measure the average number of
days worth of stock held during the period.
Inventory Days =
=
216
MARKETING METRICS
Days in Year (365)
Inventory Turns
365
5
= 73 Days Worth of Inventory
Data Sources, Complications, and Cautions
Although some companies and supply chains maintain sophisticated inventory tracking
systems, others must estimate logistical metrics on the basis of less-than-perfect data.
Increasingly, manufacturers may also have difficulty purchasing research because retailers that gather such information tend to restrict access or charge high fees for it. Often,
the only readily available data may be drawn from incomplete store audits or reports
filed by an overloaded sales force. Ideally, marketers would like to have reliable metrics
for the following:
■
■
■
■
Inventory units and monetary value of each SKU at each level of the distribution chain for each major customer.
Out-of-stocks for each SKU, measured at both the supplier and the store
level.
Percentage of customer orders that were delivered on time and in the correct
amount.
Inventory counts in the tracking system that don’t match the number in the
physical inventory. (This would facilitate a measure of shrinkage or theft.)
When considering the monetary value of inventory, it is important to use comparable
figures in all calculations. As an example of the inconsistency and confusion that can
arise in this area, a company might value its stock on the retail shelf at the cost to the
store, which might include an approximation of all direct costs. Or it might value that
stock for some purposes at the retail price. Such figures can be difficult to reconcile with
the cost of goods purchased at the warehouse and can also be different from accounting
figures adjusted for obsolescence.
When evaluating inventory, managers must also establish a costing system for items
that can’t be tracked on an individual basis. Such systems include the following:
■
■
First In, First Out (FIFO): The first unit of inventory received is the first
expensed upon sale.
Last In, First Out (LIFO): The last unit of inventory received is the first
expensed upon sale.
The choice of FIFO or LIFO can have a significant financial impact in inflationary
times. At such times, FIFO will hold down the cost of goods sold by reporting this figure at the earliest available prices. Simultaneously, it will value inventory at its highest
possible level—that is, at the most recent prices. The financial impact of LIFO will be
the reverse.
In some industries, inventory management is a core skill. Examples include the apparel
industry, in which retailers must ensure that they are not left with prior seasons’
Chapter 6 Sales Force and Channel Management
217
fashions, and the technology industry, in which rapid developments make products
hard to sell after only a few months.
In logistical management, firms must beware of creating reward structures that lead to
sub-optimal outcomes. An inventory manager rewarded solely for minimizing out-ofstocks, for example, would have a clear incentive to overbuy—regardless of inventory
holding costs. In this field, managers must ensure that incentive systems are sophisticated enough not to reward undesirable behavior.
Firms must also be realistic about what will be achieved in inventory management. In
most organizations, the only way to be completely in stock on every product all the time
is to ramp up inventories. This will involve huge warehousing costs. It will tie up a great
deal of the company’s capital in buying stocks. And it will result in painful obsolescence
charges to unload over-purchased items. Good logistics and inventory management
entails finding the right trade-off between two conflicting objectives: minimizing both
inventory holding costs and sales lost due to out-of-stocks.
Related Metrics and Concepts
Rain Checks, or Make-Goods on Promotions: These measures evaluate the effect
on a store of promotional items being unavailable. In a typical example, a store might
track the incidents in which it offers customers a substitute item because it has run
out of stock on a promoted item. Rain checks or make-goods might be expressed as a
percentage of goods sold, or more specifically, as a percentage of revenues coded to the
promotion but generated by sales of items not listed as part of the promotional event.
Misshipments: This measures the number of shipments that failed to arrive on time or
in the proper quantities.
Deductions: This measures the value of deductions from customer invoices caused by
incorrect or incomplete shipments, damaged goods, returns, or other factors. It is often
useful to distinguish between the reasons for deductions.
Obsolescence: This is a vital metric for many retailers, especially those involved in
fashion and technology. It is typically expressed as the monetary value of items that
are obsolete, or as the percentage of total stock value that comprises obsolete items. If
obsolescence is high, then a firm holds a significant amount of inventory that is likely
to sell only at a considerable discount.
Shrinkage: This is generally a euphemism for theft. It describes a phenomenon in
which the value of actual inventory runs lower than recorded inventory, due to an
unexplained reduction in the number of units held. This measure is typically calculated
as a monetary figure or as a percentage of total stock value.
Pipeline Sales: Sales that are required to supply retail and wholesale channels with sufficient inventory to make a product available for sale (refer to Section 6.5).
218
MARKETING METRICS
Consumer Off-Take: Purchases by consumers from retailers, as opposed to purchases
by retailers or wholesalers from their suppliers. When consumer off-take runs higher
than manufacturer sales rates, inventories will be drawn down.
Diverted Merchandise or Diverted Goods: Products shipped to one customer that are
subsequently resold to another customer. For example, if a retail drug chain overbuys
vitamins at a promotional price, it may ship some of its excess inventory to a dollar store.
6.8 SKU Profitability: Markdowns, GMROII, and DPP
Profitability metrics for retail products and categories are generally similar to other
measures of profitability, such as unit and percentage margins. Certain refinements have been developed for retailers and distributors, however. Markdowns,
for example, are calculated as a ratio of discount to original price charged. Gross
margin return on inventory investment (GMROII) is calculated as margin divided
by the cost of inventory and is expressed as a “rate” or percentage. Direct product
profitability (DPP) is a metric that adjusts gross margin for other costs, such as
storage, handling, and allowances paid by suppliers.
Markdown (%) =
Reduction in Price of SKU ($)
Initial Price of SKU ($)
Gross Margin Return on Inventory Investment (%) =
Gross Margin on Product
Sales in Period ($)
Average Inventory Value at
Cost ($)
Direct Product Profitability ($) = Gross Margin ($) – Direct Product Costs ($)
By monitoring markdowns, marketers can gain important insight into SKU
profitability. GMROII can be a vital metric in determining whether sales rates justify inventory positions. DPP is a theoretically powerful measure of profit that has
fallen out of favor, but it may be revived in other forms (for example, activity-based
costing).
Purpose: To assess the effectiveness and profitability of individual
product and category sales.
Retailers and distributors have a great deal of choice regarding which products to stock
and which to discontinue as they make room for a steady stream of new offerings. By
measuring the profitability of individual stock keeping units (SKUs), managers develop
Chapter 6 Sales Force and Channel Management
219
the insight needed to optimize such product selections. Profitability metrics are also
useful in decisions regarding pricing, display, and promotional campaigns.
Figures that affect or reflect retail profitability include markdowns, gross margin return
on inventory investment, and direct product profitability. Taking each in turn:
Markdowns are not always applied to slow-moving merchandise. Markdowns in excess
of budget, however, are almost always regarded as indicators of errors in product
assortment, pricing, or promotion. Markdowns are often expressed as a percentage of
regular price. As a standalone metric, a markdown is difficult to interpret.
Gross margin return on inventory investment (GMROII) applies the concept of return
on investment (ROI) to what is often the most crucial element of a retailer’s working
capital: its inventory.
Direct product profitability (DPP) shares many features with activity-based costing
(ABC). Under ABC, a wide range of costs are weighted and allocated to specific products
through cost drivers—the factors that cause the costs to be incurred. In measuring DPP,
retailers factor such line items as storage, handling, manufacturer’s allowances, warranties, and financing plans into calculations of earnings on specific product sales.
Construction
Markdown: This metric quantifies shop-floor reductions in the price of a SKU. It can
be expressed on a per-unit basis or as a total for the SKU. It can also be calculated in
dollar terms or as a percentage of the item’s initial price.
Markdown ($) = Initial Price of SKU ($) – Actual Sales Price ($)
Markdown (%) =
Markdown ($)
Initial Price of SKU ($)
Gross Margin Return on Inventory Investment (GMROII): This metric quantifies the
profitability of products in relation to the inventory investment required to make them
available. It is calculated by dividing the gross margin on product sales by the cost of
the relevant inventory.
Gross Margin Return on Inventory Investment (%) =
220
MARKETING METRICS
Gross Margin on Product Sales in
Period ($)
Average Inventory Value at Cost ($)
Direct Product Profitability (DPP)
Direct product profitability is grounded in a simple concept, but it can be difficult to
measure in practice. The calculation of DPP consists of multiple stages. The first stage
is to determine the gross margin of the goods in question. This gross margin figure is
then modified to take account of other revenues associated with the product, such as
promotional rebates from suppliers or payments from financing companies that gain
business on its sale. The adjusted gross margin is then reduced by an allocation of direct
product costs, described next.
Direct Product Costs: These are the costs of bringing a product to customers. They
generally include warehouse, distribution, and store costs.
Direct Product Costs ($) = Warehouse Direct Costs ($) + Transportation Direct Costs ($)
+ Store Direct Costs ($)
Direct Product Profitability (DPP): Direct product profitability represents a product’s
adjusted gross margin, less its direct product costs.
As noted earlier, the concept of DPP is quite simple. Difficulties can arise, however,
in calculating or estimating the relevant costs. Typically, an elaborate ABC system is
needed to generate direct costs for individual SKUs. DPP has fallen somewhat out of
favor as a result of these difficulties.
Other metrics have been developed, however, in an effort to obtain a more refined and
accurate estimation of the “true” profitability of individual SKUs, factoring in the varying costs of receiving, storing, and selling them. The variations between products in the
levels of these costs can be quite significant. In the grocery industry, for example, the
cost of warehousing and shelving frozen foods is far greater—per unit or per dollar of
sales—than the cost of warehousing and shelving canned goods.
Direct Product Profitability ($) = Gross Margin ($) – Direct Product Costs ($)
EXAMPLE: The apparel retailer cited earlier wants to probe further into the profitability of its
sock line. Toward that end, it assembles the following information. For this retailer, socks generate slotting allowances—in essence, fees paid by the manufacturer to the retailer in compensation for shelf space—in the amount of $50,000 per year. Warehouse costs for the retailer come to
$10,000,000 per year. Socks consume 0.5% of warehouse space. Estimated store and distribution
costs associated with socks total $80,000.
With this information, the retailer calculates an adjusted gross margin for its sock line.
Adjusted Gross Margin = Gross Margin + Additional Margin
= $350,000 + $50,000
= $400,000
Chapter 6 Sales Force and Channel Management
221
The retailer then calculates direct product costs for its sock line.
Direct Product Costs = Store and Distribution Costs + Warehouse Costs
= $80,000 + (0.5% * $10,000,000)
= $80,000 + $50,000
= $130,000
On this basis, the retailer calculates the direct product profitability of its sock line.
DPP = Gross Margin – Direct Product Costs
= $400,000 – $130,000
= $270,000
Data Sources, Complications, and Cautions
For GMROII calculations, it is necessary to determine the value of inventory held, at
cost. Ideally, this will be an average figure for the period to be considered. The average
of inventory held at the beginning and end of the period is often used as a proxy, and
is generally—but not always—an acceptable approximation. To perform the GMROII
calculation, it is also necessary to calculate a gross margin figure.
One of the central considerations in evaluating direct product profitability is an organization’s ability to capture large amounts of accurate data for analysis. The DPP calculation requires an estimate of the warehousing, distribution, store direct, and other
costs attributable to a product. To assemble these data, it may be necessary to gather all
distribution costs and apportion them according to the cost drivers identified.
Inventory held, and thus the cost of holding it, can change considerably over time.
Although one may usually approximate average inventory over a period by averaging the beginning and ending levels of this line item, this will not always be the case.
Seasonal factors may perturb these figures. Also, a firm may hold substantially more—
or less—inventory during the course of a year than at its beginning and end. This could
have a major impact on any DPP calculation.
DPP also requires a measure of the ancillary revenues tied to product sales.
Direct product profitability has great conceptual strength. It tries to account for the
wide range of costs that retailers incur in conveying a product to customers, and thus to
yield a more realistic measure of the profitability of that product. The only significant
weakness in this metric is its complexity. Few retailers have been able to implement it.
Many firms continue to try to realize its underlying concept, however, through such
programs as activity-based costing.
222
MARKETING METRICS
Related Metrics and Concepts
Shopping Basket Margin: The profit margin on an entire retail transaction, which may
include a number of products. This aggregate transaction is termed the “basket” of purchases that a consumer makes.
One key factor in a firm’s profitability is its capability to sell ancillary products in addition to its central offering. In some businesses, more profit can be generated through
accessories than through the core product. Beverage and snack sales at movie theaters
are a prime example. With this in mind, marketers must understand each product’s role
within their firm’s aggregate offering—be it a vehicle to generate customer traffic, or to
increase the size of each customer’s basket, or to maximize earnings on that item itself.
References and Suggested Further Reading
Wilner, J.D. (1998). 7 Secrets to Successful Sales Management: The Sales Manager’s Manual, Boca
Raton: St. Lucie Press.
Zoltners, A.A., P. Sinha, and G.A. Zoltners. (2001). The Complete Guide to Accelerating Sales
Force Performance, New York: Amacom.
Chapter 6 Sales Force and Channel Management
223
This page intentionally left blank
7
PRICING STRATEGY
Introduction
Key concepts covered in this chapter:
Price Premium
Reservation Price
Percent Good Value
Optimal Prices, Linear and Constant
Demand
“Own,” “Cross,” and Price Elasticity
Price Elasticity of Demand
“The cost of . . . lack of sophistication in pricing is growing day by day. Customers
and Competitors operating globally in a generally more complex marketing environment are making mundane thinking about pricing a serious threat to the firm’s
financial well being.”1
A full-fledged evaluation of pricing strategies and tactics is well beyond the scope of this
book. However, there are certain key metrics and concepts that are fundamental to the
analysis of pricing alternatives, and this chapter addresses them.
First we describe several of the more common methods of calculating price premiums—
also called relative prices.
Next, we discuss the concepts that form the foundation of price-quantity schedules—
also known as demand functions or demand curves. These include reservation prices
and percent good value.
In the third section, we explain the definition and calculation of price elasticity, a frequently used index of market response to changes in price. This relatively simple ratio
225
of percentage changes in volumes and prices is complicated in practice by variations in
measure and interpretation.
For managers, the purpose of understanding price elasticity is to improve pricing. With
this in mind, we’ve devoted a separate section to determining optimal prices for the two
main types of demand functions: linear and constant elasticity. The final portion of this
chapter addresses the question of whether elasticity has been calculated in a manner
that incorporates likely competitive reactions. It explains three types of elasticity—
“own,” “cross,” and “residual” elasticity. Although these may seem at first glance to
rest upon subtle or pedantic distinctions, they have major pragmatic implications. The
familiar concept of the prisoner’s dilemma helps explain their import.
Metric
7.1
Construction
Price Premium The percentage by
which the price of
a brand exceeds a
benchmark price.
Considerations
Purpose
Benchmarks
include average
price paid, average
price charged,
average price
displayed, and price of
a relevant competitor.
Prices can be
compared at any level
in the channel
and can be calculated
on a gross basis or net
of discounts and
rebates.
Measures how
a brand’s price
compares to that of
its competition.
7.2
Reservation
Price
The maximum
amount an
individual is
willing to pay for a
product.
Reservation prices are
difficult to observe.
One way to
conceptualize a
demand curve is
as the aggregation
of reservation
prices of potential
customers.
7.2
Percent Good
Value
The proportion
of customers
who consider a
product to be a
good value—that
is, to have a selling
price below their
reservation price.
Easier to observe than
individual reservation
prices.
A second way to
conceptualize a
demand curve is
as the relationship
between percent
good value and
price.
226
MARKETING METRICS
Metric
Construction
Considerations
Purpose
7.3
Price Elasticity The responsiveness
of Demand
of demand to a
small change in
price, expressed
as a ratio of
percentages.
For linear demand,
linear projections
based on elasticity
are accurate, but
elasticity changes with
price. For constant
elasticity demand,
linear projections
are approximate, but
elasticity is the same
for all prices.
Measures the
responsiveness
of quantity to
changes in price. If
priced optimally,
the margin is the
negative inverse of
elasticity.
7.4
Optimal Price
For linear demand,
optimal price is
the average of
variable cost and
the maximum
reservation price.
For constant
elasticity, optimal
price is a known
function of variable
cost and elasticity.
In general, optimal
price is the price
that maximizes
contribution after
accounting for how
quantity changes
with price.
Optimal price formulas
are appropriate only
if the variable cost
per unit is constant,
and there are no
larger strategic
considerations.
Quickly determines
the price that
maximizes
contribution.
7.5
Residual
Elasticity
Residual elasticity
is “own” elasticity
plus the product of
competitor reaction
elasticity and cross
elasticity.
Rests on an
assumption that
competitor reaction to
a firm’s price changes
is predictable.
Measures the
responsiveness
of quantity to
changes in price,
after accounting
for competitor
reactions.
Chapter 7 Pricing Strategy
227
7.1 Price Premium
Price premium, or relative price, is the percentage by which a product’s selling price
exceeds (or falls short of) a benchmark price.
Price Premium (%) =
Brand A Price ($) – Benchmark Price ($)
Benchmark Price ($)
Marketers need to monitor price premiums as early indicators of competitive pricing
strategies. Changes in price premiums can also be signs of product shortages, excess
inventories, or other changes in the relationships between supply and demand.
Purpose: To evaluate product pricing in the context
of market competition.
Although there are several useful benchmarks with which a manager can compare a
brand’s price, they all attempt to measure the “average price” in the marketplace. By
comparing a brand’s price with a market average, managers can gain valuable insight
into its strength, especially if they view these findings in the context of volume and
market share changes. Indeed, price premium—also known as relative price—is a commonly used metric among marketers and senior managers. Fully 63% of firms report
the Relative Prices of their products to their boards, according to a recent survey conducted in the U.S., UK, Germany, Japan, and France.2
Price Premium: The percentage by which the price charged for a specified brand
exceeds (or falls short of) a benchmark price established for a similar product or
basket of products. Price premium is also known as relative price.
Construction
In calculating price premium, managers must first specify a benchmark price. Typically,
the price of the brand in question will be included in this benchmark, and all prices
in the benchmark will be for an equivalent volume of product (for example, price per
liter). There are at least four commonly used benchmarks:
■
The price of a specified competitor or competitors.
■
Average price paid: The unit-sales weighted average price in the category.
■
Average price displayed: The display-weighted average price in the category.
■
Average price charged: The simple (unweighted) average price in the category.
Price of a Specified Competitor: The simplest calculation of price premium
involves the comparison of a brand’s price to that of a direct competitor.
228
MARKETING METRICS
EXAMPLE: Ali’s company sells “gO2” mineral water in its EU home market at a 12% premium
over the price of its main competitor. Ali would like to know whether the same price premium is
being maintained in the Turkish market, where gO2 faces quite different competition. He notes
that gO2 mineral water sells in Turkey for 2 (new) Lira per liter, while its main competitor, Essence,
sells for 1.9 Lira per liter.
Price Premium =
=
(2.0 YTL – 1.9 YTL)
1.9 YTL
0.1 YTL
1.9 YTL
= 5.3% Premium Versus Essence
When assessing a brand’s price premium vis à vis multiple competitors, managers can
use as their benchmark the average price of a selected group of those competitors.
Average Price Paid: Another useful benchmark is the average price that customers pay for brands in a given category. This average can be calculated in at least
two ways: (1) as the ratio of total category revenue to total category unit sales, or
(2) as the unit-share weighted average price in the category. Note that the market
Average Price Paid includes the brand under consideration.
Note also that changes in unit shares will affect the average price paid. If a low-price
brand steals shares from a higher-priced rival, the average price paid will decline. This
would cause a firm’s price premium (calculated using the average price paid as a benchmark) to rise, even if its absolute price did not change. Similarly, if a brand is priced at a
premium, that premium will decline as it gains share. The reason: A market share gain by a
premium-priced brand will cause the overall average price paid in its market to rise. This,
in turn, will reduce the price differential between that brand and the market average.
EXAMPLE: Ali wants to compare his brand’s price to the average price paid for similar products in the market. He notes that gO2 sells for 2.0 Lira per liter and has 20% of the unit sales in
market. Its up-market competitor, Panache, sells for 2.1 Lira and enjoys 10% unit market share.
Essence sells for 1.9 Lira and has 20% share. Finally, the budget brand, Besik, sells for 1.2 Lira and
commands 50% of the market.
Ali calculates the weighted Average Price Paid as (20% * 2) + (10% * 2.1)
+ (20% * 1.9) + (50% * 1.2) = 1.59 Lira.
Price Premium (%) =
=
(2.00 – 1.59)
1.59
0.41
1.59
= 25.8%
Chapter 7 Pricing Strategy
229
To calculate the price premium using the average price paid benchmark, managers can
also divide a brand’s share of the market in value terms by its share in volume terms. If
value and volume market shares are equal, there is no premium. If value share is greater
than volume share, then there is a positive price premium.
Revenue Market Share (%)
Price Premium (%) =
Unit Market Share (%)
Average Price Charged: Calculation of the average price paid requires knowledge
of the sales or shares of each competitor. A much simpler benchmark is the average
price charged—the simple unweighted average price of the brands in the category.
This benchmark requires knowledge only of prices. As a consequence, the price
premium calculated using this benchmark is not affected by changes in unit shares.
For this reason, this benchmark serves a slightly different purpose. It captures the
way a brand’s price compares to prices set by its competitors, without regard to
customers’ reactions to those prices. It also treats all competitors equally in the calculation of the benchmark price. Large and small competitors are weighted equally
when calculating average price charged.
EXAMPLE: Using the previous data, Ali also calculates the average price charged in the mineral water category as (2 + 2.1 + 1.9 + 1.2)/4 = 1.8 Lira.
Using the average price charged as his benchmark, he calculates gO2’s price premium as
Price Premium (%) =
=
(2.0 – 1.8)
1.8
0.2
1.8
= 11.1% Premium
Average Price Displayed: One benchmark conceptually situated between average price paid and average price charged is the average price displayed. Marketing
managers who seek a benchmark that captures differences in the scale and strength
of brands’ distribution might weight each brand’s price in proportion to a numerical measure of distribution. Typical measures of distribution strength include
numeric distribution, ACV (%), and PCV (%).
EXAMPLE:
Ali calculates the average price displayed using numeric distribution.
Ali’s brand, gO2, is priced at 2 Lira and is distributed in 500 of the 1,000 stores that carry bottled
water. Panache is priced at 2.1 Lira and stocked by 200 stores. Essence is priced at 1.9 Lira and sold
through 400 stores. Besik carries a price of 1.2 Lira and has a presence in 900 stores.
230
MARKETING METRICS
Ali calculates relative weighting on the basis of numeric distribution. The total number of stores is
1,000. The weightings are therefore, for gO2, 500/1,000 = 50%; for Panache, 200/1,000 = 20%; for
Essence, 400/1,000 = 40%; and for Besik, 900/1,000 = 90%. As the weightings thus total 200%, in
calculating average price displayed, the sum of the weighted prices must be divided by that figure,
as follows:
Average Price Displayed =
[(2 * 50%) + (2.1 * 20%) + (1.9 * 40%) + (1.2 * 90%)]
200%
= 1.63 Lira
Price Premium (%) =
=
(2.00 – 1.63)
1.63
0.37
1.63
= 22.7% premium
Data Sources, Complications, and Cautions
There are several practical aspects of calculating price premiums that deserve mention.
Managers may find it easier to select a few leading competitors and focus their analysis and
comparison on these. Often, it is difficult to obtain reliable data on smaller competitors.
Managers must exercise care when interpreting price premiums. Different benchmarks
measure different types of premiums and must be interpreted accordingly.
Can a price premium be negative? Yes. Although generally expressed in terms that
imply only positive values, a price premium can be negative. If one brand doesn’t command a positive premium, a competitor will. Consequently, except in the unlikely event
that all prices are exactly equal, managers may want to speak in terms of positive premiums. When a given brand’s price is at the low end of the market, managers may want to
say that the competition holds a price premium of a certain value.
Should we use retail, manufacturer, or distributor pricing? Each is useful in understanding the market dynamics at its level. When products have different channel
margins, their price premiums will differ, depending on the channel under consideration. When stating a price premium, managers are advised to specify the level to which
it applies.
Prices at each level can be calculated on a gross basis, or net of discounts, rebates,
and coupons. Especially when dealing with distributors or retailers, there are likely to
be substantial differences between manufacturer selling prices (retail purchase prices),
depending on whether they are adjusted for discounts and allowances.
Chapter 7 Pricing Strategy
231
Related Metrics and Concepts
Theoretical Price Premium: This is the price difference that would make potential customers indifferent between two competing products. It represents a different use of the
term “price premium” that is growing in popularity. The theoretical price premium can
also be discovered through a conjoint analysis using brand as an attribute. The theoretical price premium is the point at which consumers would be indifferent between a
branded and an unbranded item, or between two different brands. We have termed this
a “theoretical” price premium because there is no guarantee that the price premiums
observed in the market will take this value. (Refer to Section 4.5 for an explanation of
conjoint analysis.)
7.2 Reservation Price and Percent Good Value
The reservation price is the value a customer places on a product. It constitutes an
individual’s maximum willingness to pay. Percent good value represents the proportion of customers who believe a product is a “good value” at a specific price.
These are useful metrics in marketers’ evaluation of pricing and customer value.
Purpose
Reservation prices provide a basis for estimating products’ demand functions in situations where other data are not available. They also offer marketers insight into pricing
latitude. When it is not possible or convenient to ask customers about their reservation
prices, percent good value can provide a substitute for that metric.
Construction
Reservation Price: The price above which a customer will not buy a product. Also
known as the maximum willingness to pay.
Percent Good Value: The proportion of customers who perceive a product to
represent a good value, that is, to carry a selling price at or below their reservation
price.
By way of example, let’s posit a market consisting of 11 individuals with reservation
prices for a given product of $30, $40, $50, $60, $70, $80, $90, $100, $110, $120, and
$130. The manufacturer of that product seeks to decide upon its price. Clearly, it might
232
MARKETING METRICS
do better than to offer a single price. For now, however, let’s assume tailored prices are
impractical. The variable cost to produce the product is $60 per unit.
With these reservation prices, the manufacturer might expect to sell 11 units at $30 or
less, 10 units at a price greater than $30 but less than or equal to $40, and so on. It would
make no sales at a unit price greater than $130. (For convenience, we have assumed that
people buy at their reservation price. This assumption is consistent with a reservation
price being the maximum an individual is willing to pay.)
Table 7.1 shows this price-quantity relationship, together with the contribution to the
firm at each possible price.
Table 7.1
Price-Quantity Relationship
Price
% Good Value
Quantity
Total Contribution
$20
100.00%
11
–$440
$30
100.00%
11
–$330
$40
90.91%
10
–$200
$50
81.82%
9
–$90
$60
72.73%
8
$0
$70
63.64%
7
$70
$80
54.55%
6
$120
$90
45.45%
5
$150
$100
36.36%
4
$160
$110
27.27%
3
$150
$120
18.18%
2
$120
$130
9.09%
1
$70
$140
0.00%
0
$0
$150
0.00%
0
$0
Variable Cost is $60 per unit.
A table of quantities expected at each of several prices is often called a demand schedule (or curve). This example shows that one way to conceptualize a demand curve is
as the accumulation of individual reservation prices. Although it will clearly be difficult in practice to measure individual reservation prices, the point here is simply
to illustrate the use of reservation prices in pricing decisions. In this example, the
Chapter 7 Pricing Strategy
233
optimal price—that is, the price that maximizes total contribution—is $100. At $100,
the manufacturer expects to sell four units. Its contribution margin is $40, yielding a
total contribution of $160.
This example also illustrates the concept of consumer surplus. At $100, the manufacturer sells three items at a price point below customers’ reservation prices. The
consumer with the reservation price of $110 enjoys a surplus of $10. The consumer
with the reservation price of $120 receives a surplus of $20. Finally, the consumer with
the highest reservation price, $130, receives a surplus of $30. From the manufacturer’s
perspective, the total consumer surplus—$60—represents an opportunity for increased
contribution if it can find a way to capture this unclaimed value.
Data Sources, Complications, and Cautions
Finding reservation prices is no easy matter. Two techniques that are frequently used to
gain insight into this metric are as follows:
■
■
Second-price auctions: In a second-price auction, the highest bidder wins
but pays only the second-highest bid amount. Auction theory suggests that
when bidding on items of known value in such auctions, individuals have an
incentive to bid their reservation prices. Certain survey techniques have been
designed to mimic this process. In one of these, customers are asked to name
their prices for an item, with the understanding that these prices will then be
subjected to a lottery. If the price drawn in the lottery is less than the price
named, the respondent gains an opportunity to purchase the item in question
at the drawn price.
Conjoint analysis: In this analytical technique, marketers gain insight into
customer perceptions regarding the value of any set of attributes through the
trade-offs they are willing to make.
Such tests can, however, be difficult to construct and impractical in many circumstances. Consequently, as a fallback technique, marketers can measure percent good
value. Rather than seeking to learn each customer’s reservation price, they may find it
easier to test a few candidate prices by asking customers whether they consider an item
a “good value” at each of those prices.
Linear Demand
The quantity-price schedule formed by an accumulation of reservation prices can take a
variety of shapes. When the distribution of reservation prices is uniform—when reservation prices are equally spaced, as in our example—the demand schedule will be linear (see
Figure 7.1). That is, each increment in price will reduce quantity by an equal amount.
As the linear function is by far the most commonly used representation of demand, we
234
MARKETING METRICS
Quantity
Demanded
Maximum Willing to Buy (MWB)
Two Points on the Linear
Demand Function
Maximum Reservation
Price (MRP)
Price
Variable Cost
Figure 7.1 Maximum Willing to Buy and Maximum Reservation Price
provide a description of this function as it relates to the distribution of underlying reservation prices.
It takes only two points to determine a straight line. Likewise, it takes only two parameters to write an equation for that line. Generally, that equation is written as Y = mX +
b, in which m is the slope of the line and b is its Y-intercept.
A line, however, can also be defined in terms of the two points where it crosses the axes.
In the case of linear demand, these crossing points (intercepts) have useful managerial
interpretations.
The quantity-axis intercept can be viewed as a representation of the maximum willing to buy (MWB). This is the total number of potential customers for a product. A
firm can serve all these customers only at a price of zero. Assuming that each potential
customer buys one unit, MWB is the quantity sold when the price is zero.
The price-axis intercept can be viewed as the maximum reservation price (MRP). The
MRP is a number slightly greater than the highest reservation price among all those
willing to buy. If a firm prices its product at or above MRP, no one will buy.
Maximum Reservation Price: The lowest price at which quantity demanded
equals zero.
Chapter 7 Pricing Strategy
235
Maximum Willing to Buy (MWB): The quantity that customers will “buy” when
the price of a product is zero. This is an artificial concept used to anchor a linear
demand function.
In a linear demand curve defined by MWB and MRP, the equation for quantity (Q) as
a function of price (P) can be written as follows:
Q ⫽ (MWB) * [1 ⫺
P
]
MRP
EXAMPLE: Erin knows that the demand for her soft drink is a simple linear function of price.
She can sell 10 units at a price of zero. When the price hits $5 per unit, demand falls to zero. How
many units will Erin sell if the price is $3 (see Figure 7.2)?
Linear Demand: Price and Quantity Demanded
12
Maximum Willing to Buy
Quantity Demanded
10
8
Maximum Reservation Price
6
4
2
0
$0
$1
$2
$3
$4
$5
Price
Figure 7.2 Simple Linear Demand (Price-Quantity) Function
For Erin’s soft drink, the MRP (Maximum Reservation Price) is $5 and the MWB (Maximum
Willing to Buy ) is 10 units. At a price of $3, Erin will sell 10 * (1 – $3/$5), or 4 units.
236
MARKETING METRICS
When demand is linear, any two points on the price-quantity demand function can be
used to determine MRP and MWB. If P1 and Q1 represent the first price-quantity point
on the line, and P2 and Q2 represent the second, then the following two equations can
be used to calculate MWB and MRP.
MWB ⫽ Q1 ⫺ (
Q 2 ⫺ Q1
) P
P 2 ⫺ P1 * 1
MRP ⫽ P1 ⫺ (
P2 ⫺ P1
)
Q2 ⫺ Q1
EXAMPLE: Early in this chapter, we met a firm that sells five units at a price of $90 and three
units at a price of $110. If demand is linear, what are MWB and MRP?
MWB = 5 – (–2/$20) * $90
=5+9
= 14
MRP = $90 – ($20/–2) * 5
= $90 + $50
= $140
The equation for quantity as a function of price is thus:
Q ⫽ 14 * (1 ⫺
P
)
$140
The market in this example, as you may recall, comprises 11 potential buyers with reservation
prices of $30, $40, . . . , $120, $130. At a price of $130, the firm sells one unit. If we set price equal to
$130 in the previous equation, our calculation does indeed result in a quantity of one. For this to
hold true, the MRP must be a number slightly higher than $130.
A linear demand function often yields a reasonable approximation of actual demand only over
a limited range of prices. In our 11-person market, for example, demand is linear only for prices
between $30 and $130. To write the equation of the linear function that describes demand between
$30 and $130, however, we must use an MWB of 14 and an MRP of $140. When we use this linear
equation, we must remember that it reflects actual demand only for prices between $30 and $130,
as illustrated in Figure 7.3.
Chapter 7 Pricing Strategy
237
Quantity
Demanded
Linear Demand Assumption
15
14
Quantity
Demanded
13
12
Linear
Demand
11
10
9
8
7
6
5
4
3
2
1
0
0
$0 $1
0
$2
0
$3
0
$4
0
$5
0
$6
0
$7
0
$8
0
$9
00
$1
10
$1
20
$1
30
$1
40
$1
Price
Figure 7.3
Example of Linear Demand Function
7.3 Price Elasticity of Demand
Price elasticity measures the responsiveness of quantity demanded to a small change
in price.
Price Elasticity (I) =
Change in Quantity (%)
Change in Price (%)
Price elasticity can be a valuable tool, enabling marketers to set an optimal price.
238
MARKETING METRICS
Purpose: To understand market responsiveness to changes in price.
Price elasticity is the most commonly employed measure of market responsiveness to
changes in price. Many marketers, however, use this term without a clear understanding of what it entails. This section will help clarify some of the potentially dangerous
details associated with estimates of price elasticity. This is challenging material but is
well worth the effort. A strong command of price elasticity can help managers set optimal prices.
Price Elasticity: The responsiveness of demand to a small change in price,
expressed as a ratio of percentages. If price elasticity is estimated at –1.5, for
example, then we expect the percentage change in quantity to be approximately
1.5 times the percentage change in price. The fact that this number is negative
indicates that when price rises, the quantity demanded is expected to decline, and
vice versa.
Construction
If we raise the price of a product, do we expect demand to hold steady or crash through
the floor? In markets that are unresponsive to price changes, we say demand is inelastic.
If minor price changes have a major impact on demand, we say demand is elastic. Most
of us have no trouble understanding elasticity at a qualitative level. The challenges come
when we quantify this important concept.
Challenge One: Questions of sign.
The first challenge in elasticity is to agree on its sign. Elasticity is the ratio of the percentage change in quantity demanded to the percentage change in price, for a small
change in price. If an increase in price leads to a decrease in quantity, this ratio will be
negative. Consequently, by this definition, elasticity will almost always be a negative
number.
Many people, however, simply assume that quantity goes down as price goes up, and
jump immediately to the question of “by how much.” For such people, price elasticity
answers that question and is a positive number. In their eyes, if elasticity is 2, then a
small percentage increase in price will yield twice that percentage decrease in quantity.
In this book, under that scenario, we would say price elasticity is –2.
Challenge Two: When demand is linear, elasticity changes with price.
For a linear demand function, the slope is constant, but elasticity is not. The reason:
Elasticity is not the same as slope. Slope is the change in quantity for a small change in
price. Elasticity, by contrast, is the percentage change in quantity for a small percentage
change in price.
Chapter 7 Pricing Strategy
239
EXAMPLE: Consider three points on a linear demand curve: ($8, 100 units), ($9, 80 units), and
($10, 60 units) (see Figure 7.4). Each dollar change in price yields a 20-unit change in quantity. The
slope of this curve is a constant –20 units per dollar.
As price rises from $8 to $9 (a 12.5% increase), quantity declines from 100 to 80 (a 20% decrease).
The ratio of these percentages is 20%/12.5%, or –1.6. Similarly, as price rises from $8 to $10 (a 25%
increase), quantity declines from 100 to 60 (a 40% decrease). Once again, the ratio (40%/25%) is
–1.6. It appears that the ratio of percentage change in quantity to percentage change in price is –1.6,
regardless of the size of the change made in the $8 price.
Linear Demand
140
120
Quantity
100
80
60
40
20
0
$6.00
$8.00
$10.00
$12.00
Price
Figure 7.4
Linear Demand Function
Consider, however, what happens when price rises from $9 to $10 (an 11.11% increase). Quantity
declines from 80 to 60 (a 25% decrease). The ratio of these figures, 25%/ 11.11%, is now –2.25. A
price decline from $9 to $8 also yields an elasticity ratio of –2.25. It appears that this ratio is –2.25 at
a price of $9, regardless of the direction of any change in price.
Exercise: Verify that the ratio of percentage change in quantity to percentage change in price at the
price of $10 is –3.33 for every conceivable price change.
240
MARKETING METRICS
For a linear demand curve, elasticity changes with price. As price increases, elasticity gains in magnitude. Thus, for a linear demand curve, the absolute unit change in
quantity for an absolute dollar change in price (slope) is constant, while the percentage
change in quantity for a percentage change in price (elasticity) is not. Demand becomes
more elastic—that is, elasticity becomes more negative—as price increases.
For a linear demand curve, the elasticity of demand can be calculated in at least
three ways:
Q 2 ⫺ Q1
Q1
Elasticity (P1) ⫽
P2 ⫺ P1
P1
⫽
Q2 ⫺ Q1
P1
*
Q
P2 ⫺ P1
1
⫽ Slope *
冢 冣
P
冢Q11冣
To emphasize the idea that elasticity changes with price on a linear demand curve, we
write “Elasticity (P),” reflecting the fact that elasticity is a function of price. We also use
the term “point elasticity” to cement the idea that a given elasticity applies only to a
single point on the linear demand curve.
Equivalently, because the slope of a linear demand curve represents the change in quantity for a given change in price, price elasticity for a linear demand curve is equal to
the slope, multiplied by the price, divided by the quantity. This is captured in the third
equation here.
EXAMPLE: Revisiting the demand function from earlier, we see that the slope of the curve
reflects a 20-unit decline in demand for each dollar increase in price. That is, slope equals –20.
The slope formula for elasticity can be used to verify our earlier calculations. Calculate price/quantity at each point on the curve, and multiply this by the slope to yield the price elasticity at that point
(see Table 7.2).
For example, at a price of $8, quantity sold is 100 units. Thus:
Elasticity ($8) = –20 * (8/100)
= –1.6
Chapter 7 Pricing Strategy
241
Table 7.2
Elasticities at a Point Calculated from the Slope of a Function
Price
Quantity Demanded
Price/Quantity
Slope
Price Elasticity at Point
$8.00
100
0.08
(20.00)
(1.60)
$9.00
80
0.11
(20.00)
(2.25)
$10.00
60
0.17
(20.00)
(3.33)
In a linear demand function, point elasticities can be used to predict the percentage change
in quantity to be expected for any percentage change in price.
EXAMPLE: Xavi manages the marketing of a toothpaste brand. He knows the brand follows a
linear demand function. At the current price of $3.00 per unit, his firm currently sells 60,000 units
with an elasticity of –2.5. A proposal is floated to raise the price to $3.18 per unit in order to standardize margins across brands. At $3.18, how many units would be sold?
The proposed change to $3.18 represents a 6% increase over the current $3 price. Because elasticity
is –2.5, such an increase can be expected to generate a decrease in unit sales of 2.5 * 6, or 15%. A 15%
reduction in current sales of 60,000 units would yield a new quantity of 0.85 * 60,000, or 51,000.
Constant Elasticity: Demand Curve with a Constantly Changing Slope
A second common form of function used to estimate demand entails constant elasticity.3 This form is responsible for the term “demand curve” because it is, indeed,
curved. In contrast with the linear demand function, the conditions in this scenario are
reversed: Elasticity is constant, while the slope changes at every point.
The assumption underlying a constant elasticity demand curve is that a small percentage change in price will cause the same percentage change in quantity, regardless of the
value of the initial price. That is, the rate of change in quantity versus price, expressed
as a ratio of percentages, is equal to a constant throughout the curve. That constant is
the elasticity.
In mathematical terms, in a constant elasticity demand function, slope multiplied by
price divided by quantity is equal to a constant (the elasticity) for all points along the
curve (see Figure 7.5). The constant elasticity function can also be expressed in an equation that is easily calculated in spreadsheets:
Q(P) ⫽ A * P ELAS
242
MARKETING METRICS
Constant Elasticity Function
120
101.7
Quantity Demanded
100
78.0
80
61.5
60
40
20
0
$6
$7
$8
$9
$10
$11
Price
Figure 7.5 Constant Elasticity
In this equation, ELAS is the price elasticity of demand. It is usually a negative number.
A is a scaling factor. It can be viewed as the quantity that would be sold at a price of $1
(assuming that $1 is a reasonable price for the product under consideration).
EXAMPLE: Plot a demand curve with a constant elasticity of –2.25 and a scaling factor of
10,943.1. For every point on this curve, a small percentage increase in price will yield a percentage
decrease in quantity that is 2.25 times as great. This 2.25 ratio holds, however, only for the very
smallest percentage changes in price. This is because the slope changes at every point. Using the
2.25 ratio to project the results of a finite percentage increase in price is always approximate.
The curve traced in this example should look like the constant elasticity curve in Figure 7.5.
More exact figures for demand at prices $8, $9, and $10 would be 101.669, 78.000, and 61.538 units.
In its way, constant elasticity is analogous to the continuous compounding of interest.
In a constant elasticity function, every small percentage increase in price generates the
same percentage decrease in quantity. These percentage decreases compound at a constant rate, leading to an overall percentage decrease that does not precisely equal the
continuous rate.
Chapter 7 Pricing Strategy
243
For this reason, given any two points on a constant elasticity demand curve, we can no
longer calculate elasticity using finite differences as we could when demand was linear.
Instead, we must use a more complicated formula grounded in natural logarithms:
ELAS ⫽
ln(Q2兾Q1)
ln(P2兾P1)
EXAMPLE: Taking any two points from the previous constant elasticity demand curve, we can
verify that elasticity is –2.25.
At $8, for example, the quantity is 101.669. Call these P1 and Q1.
At $9 the quantity is 78.000. Call these P2 and Q2.
Inserting these into our formula, we determine that
ln(78.000兾101.669)
ELAS ⫽
ln(9兾8)
–0.265
= 0.118
= –2.25
If we had set P2 equal to $8, and P1 equal to $9, we would have arrived at the same figure for elasticity.
In fact, regardless of which two points we select on this constant elasticity curve, and regardless of the
order in which we consider them, elasticity will always be –2.25.
In summary, elasticity is the standard measure of market responsiveness to changes in
price. In general, it is the “percentage slope” of the demand function (curve) obtained
by multiplying the slope of the curve for a given price by the ratio of price to quantity.
Elasticity(P) ⫽ Slope *
冢QP 冣
Elasticity can also be viewed as the percentage change in quantity for a small percentage
change in price.
In a linear demand function, the slope is constant, but elasticity changes with price.
In this scenario, marketers can use elasticity estimates to calculate the result of an
anticipated price change in either direction, but they must use the elasticity that
is appropriate for their initial price point. The reason: In a linear demand function, elasticity varies across price points, but projections based on these elasticities
are accurate.
In a constant elasticity demand function, elasticity is the same at all price points, but
projections based on these elasticities will be approximate. Assuming they are estimated
244
MARKETING METRICS
with precision, using the constant elasticity demand function itself to make sales projections on the basis of price changes will be more accurate.
Data Sources, Complications, and Cautions
Price elasticity is generally estimated on the basis of available data. These data can be
drawn from actual sales and price changes observed in the market, conjoint studies of
customer intentions, consumer surveys about reservation prices or percent good value,
or test-market results. In deriving elasticity, price-quantity functions can be sketched
on paper, estimated from regressions in the form of linear or constant elasticity equations, or estimated through more complex expressions that include other variables in
the marketing mix, such as advertising or product quality.
To confirm the validity and usefulness of these procedures, marketers must thoroughly
understand the implications of the resulting elasticity estimate for customer behavior.
Through this understanding, marketers can determine whether their estimate makes
sense or requires further validation. That done, the next step is to use it to decide on
pricing.
7.4 Optimal Prices and Linear and Constant Demand
Functions
The optimal price is the most profitable price for any product. In a linear demand
function, the optimal price is halfway between the maximum reservation price and
the variable cost of the product.
Optimal Price for a Linear Demand Function ($) =
Maximum Reservation Price ($)
+ Variable Cost ($)
2
Generally, the gross margin on a product at its optimal price will be the negative
inverse of its price elasticity.
–1
Gross Margin at Optimal Price (%) =
Elasticity (I)
Although it can be difficult to apply, this relationship offers a powerful insight: In a
constant elasticity demand function, optimal margin follows directly from elasticity.
This greatly simplifies the determination of the optimal price for a product of known
variable cost.
Chapter 7 Pricing Strategy
245
Purpose: To determine the price that yields the greatest possible
contribution.
Although “optimal price” can be defined in a number of ways, a good starting point is
the price that will generate the greatest contribution by a product after deducting its
variable cost—that is, the most profitable price for the product.
If managers set price too low, they forego revenue from customers who would willingly
have paid more. In addition, a low price can lead customers to value a product less than
they otherwise might. That is, it causes them to lower their reservation prices.
By contrast, if managers set price too high, they risk losing contribution from people
who could have been served profitably.
Construction
For linear demand, the optimal price is the midpoint between the maximum reservation price and the variable cost of the product.
In linear demand functions, the price that maximizes total contribution for a product is
always precisely halfway between the maximum reservation price (MRP) and the variable cost to produce that product. Mathematically, if P* represents the optimal price of
a product, MRP is the X-intercept of its linear demand function, and VC is its variable
cost per unit:
P* ⫽ (MRP ⫹ VC)兾2
EXAMPLE: Jaime’s business sells goods that cost $1 to produce. Demand is linear. If priced at
$5, Jaime believes he won’t sell anything. For every dollar decrease in price, Jaime believes he will
sell one additional unit.
Given that the variable cost is $1, the maximum reservation price is $5, and the demand function
is linear, Jaime can anticipate that he’ll achieve maximum contribution at a price midway point
between VC and MRP. That is, the optimal price is ($5 + $1)/2, or $3.00 (see Figure 7.6).4
In a linear demand function, managers don’t need to know the quantity of a product demanded in
order to determine its optimal price. For those who seek to examine Jaime’s contribution figures,
however, please find the details in Table 7.3.
246
MARKETING METRICS
Maximum Total Contribution When “Square” is Formed
5
Total
Contribution
Quantity Demanded
4
Variable
Cost
3
2
Quantity
Demanded
Contribution
1
0
$-
$1
$2
$3
$4
$5
-1
Price/Cost
Figure 7.6
Optimal Price Midway Between Variable Cost and MRP
Table 7.3 Optimal Price = 1/2 (MRP + Variable Cost)
Price
Quantity
Demanded
Variable Cost
per Unit
Contribution
per Unit
Total
Contribution
$0
5
$1
($1)
($5)
$1
4
$1
$0
$0
$2
3
$1
$1
$3
$3
2
$1
$2
$4
$4
1
$1
$3
$3
$5
0
$1
$4
$0
Chapter 7 Pricing Strategy
247
The previous optimal price formula does not reveal the quantity sold at a given price or
the resulting contribution. To determine optimal contribution, managers can use the
following equation:
Contribution* ⫽ (MWB/MRP) * (P* ⫺ VC)2
EXAMPLE: Jaime develops a new but similar product. Its demand follows a linear function in
which the maximum willing to buy (MWB) is 200 and the maximum reservation price (MRP) is
$10. Variable cost is $1 per unit. Jaime knows that his optimal price will be midway between MRP
and variable cost. That is, it will be ($1 + $10)/2 = $5.50 per unit. Using the formula for optimal
contribution, Jaime calculates total contribution at the optimal price:
Contribution at Optimal Price for a Linear Demand Function ($)
= [MWB (#)/MRP ($)] * [Price ($) – Variable Costs ($)] ^ 2
= (200/10) * ($5.50 – $1) ^ 2
= 20 * $4.5 ^ 2
= $405
Jaime builds a spreadsheet that supports this calculation (see Table 7.4).
Table 7.4 Contribution Maximized at the Optimal Price
Price
Variable
Costs
Quantity
Demanded
Contribution
per Unit
Total
Contribution
$6
$1
80
$5.00
$400
$5.50
$1
90
$4.50
$405
$5
$1
100
$4.00
$400
$4
$1
120
$3.00
$360
$3
$1
140
$2.00
$280
$2
$1
160
$1.00
$160
$1
$1
180
$0.00
$0
248
MARKETING METRICS
This relationship holds across all linear demand functions, regardless of slope. For
such functions, it is therefore possible to calculate the optimal price for a product on
the basis of only two inputs: variable cost per unit and the maximum reservation price.
EXAMPLE: Brands A, B, and C each have a variable cost of $2 per unit and follow linear
demand functions as shown in Table 7.5.
Table 7.5 The Optimal Price Formula Applies to All Linear Demand Functions
Price
Demand Brand A
Demand Brand B
Demand Brand C
$2
12
20
16
$3
10
18
15
$4
8
16
14
$5
6
14
13
$6
4
12
12
$7
2
10
11
$8
0
8
10
$9
0
6
9
$10
0
4
8
$11
0
2
7
$12
0
0
6
On the basis of these inputs, we can determine the maximum reservation price—the
lowest price at which demand is zero. For Brand C, for example, we know that demand
follows a linear function in which quantity declines by one unit for each dollar increase
in price. If six units are demanded at $12, then $18 will be the lowest price at which no
one will buy a single unit. This is the maximum reservation price. We can make similar
determinations for Brands A and B (see Table 7.6).
Chapter 7 Pricing Strategy
249
Table 7.6 In Linear Demand Functions, the Determination of
Optimal Price Requires Only Two Inputs
Brand A
Brand B
Brand C
Maximum Reservation Price
$8
$12
$18
Variable Costs
$2
$2
$2
Optimal Price
$5
$7
$10
To verify that the optimal prices so determined will generate the maximum attainable contribution,
please see Table 7.7.
Table 7.7 The Optimal Prices for Linear Demand Functions Can Be Verified
Total
Demand ContriBrand A bution
(Given) Brand A
Demand
Brand B
(Given)
Total
Contribution
Brand B
Total
Demand ContriBrand C bution
(Given) Brand C
Q
Q*UC
Q
Q*UC
Q
Q*UC
Price
Variable
costs
Unit
Contribution
= P - VC
P
VC
UC
$2
$2
$0
12
$0
20
$0
16
$0
$3
$2
$1
10
$10
18
$18
15
$15
$4
$2
$2
8
$16
16
$32
14
$28
$5
$2
$3
6
$18
14
$42
13
$39
$6
$2
$4
4
$16
12
$48
12
$48
$7
$2
$5
2
$10
10
$50
11
$55
$8
$2
$6
0
$0
8
$48
10
$60
$9
$2
$7
0
$0
6
$42
9
$63
$10
$2
$8
0
$0
4
$32
8
$64
$11
$2
$9
0
$0
2
$18
7
$63
$12
$2
$10
$0
0
$0
6
$60
0
Because slope doesn’t influence optimal price, all demand functions with the same
maximum reservation price and variable cost will yield the same optimal price.
250
MARKETING METRICS
EXAMPLE: A manufacturer of chair cushions operates in three different markets—urban,
suburban, and rural. These vary greatly in size. Demand is far higher in the city than in the suburbs
or the country. Variable cost, however, is the same in all markets at $4 per unit. The maximum reservation price, at $20 per unit, is also the same in all markets. Regardless of market size, the optimal
price is therefore $12 per unit in all three markets (see Figure 7.7 and Table 7.8).
The optimal price of $12 is verified by the calculations in Table 7.9.
Different Linear Demand Functions Slopes with the Same
MWP and VC
35
Variable Cost $4
30
Quantity Demanded
25
Suburban
Demand
20
Urban
Demand
15
10
5
Rural Demand
0
$-
$2
$4
$6
$8
$10
$12
$14
$16
$18
$20
Price
Figure 7.7 Linear Demand Functions with the Same MRP and Variable Cost
Table 7.8 The Slope Doesn’t Influence Optimal Price
Maximum Reservation Price
$20
Variable Cost
$4
Optimal Price
$12
Chapter 7 Pricing Strategy
251
Table 7.9 Linear Demand Functions with Different Slopes
Price
Contribution
Suburban
Demand
Rural
Demand
Urban
Demand
Suburban
Contribution
Rural
Contribution
Urban
Contribution
$0
($4)
20
10
32
($80)
($40)
($128)
$2
($2)
18
9
29
($36)
($18)
($58)
$4
$0
16
8
26
$0
$0
$0
$6
$2
14
7
22
$28
$14
$45
$8
$4
12
6
19
$48
$24
$77
$10
$6
10
5
16
$60
$30
$96
$12
$8
8
4
13
$64
$32
$102
$14
$10
6
3
10
$60
$30
$96
$16
$12
4
2
6
$48
$24
$77
$18
$14
2
1
3
$28
$14
$45
$20
$16
—
—
—
—
—
—
In this example, it might help to think of the urban, suburban, and rural markets as groups of
people with identical, uniform distributions of reservation prices. In each, the reservation prices
are uniform between $0 and the maximum reservation price (MRP). The only difference between
segments is the number of people in each. That number represents the maximum willing to buy
(MWB). As might be expected, the number of people in a segment doesn’t affect optimal price as
much as the distribution of reservation prices in that segment. As all three segments here show the
same distribution of reservation prices, they all carry the same optimal price.
Another useful exercise is to consider what would happen if the manufacturer in this example were
able to increase everyone’s reservation price by $1. This would raise the optimal price by half that
amount, or $0.50. Likewise, the optimal price would rise by half the amount of any increase in variable cost.
Optimal Price in General
When demand is linear, we have an easy-to-use formula for optimal price. Regardless of
the shape of the demand function, there is a simple relationship between gross margin
and elasticity at the optimal price.
252
MARKETING METRICS
Optimal Price, Relative to Gross Margin: The optimal price is the price at which
a product’s gross margin is equal to the negative of the reciprocal of its elasticity
of demand.5
Gross Margin at Optimal Price (%) =
–1
Elasticity at Optimal Price
A relationship such as this, which holds at the optimal price, is called an optimality condition. If elasticity is constant, then we can easily use this optimality condition to determine
the optimal price. We simply find the negative of the reciprocal of the constant elasticity.
The result will be the optimal gross margin. If variable costs are known and constant, then
we need only determine the price that corresponds to the calculated optimal margin.
EXAMPLE: The manager of a stall selling replica sporting goods knows that the demand for
jerseys has a constant price elasticity of –4. To price optimally, she sets her gross margin equal to
the negative of the reciprocal of the elasticity of demand. (Some economists refer to the price-cost
margin as the Lerner Index.)
Gross Margin at Optimal Price =
–1
–4
= 25%
If the variable cost of each jersey is $5, the optimal price will be $5/(1 – 0.25), or $6.67.
The optimal margins for several price elasticities are listed in Table 7.10.
Table 7.10
Optimal Margins for Sample Elasticities
Price Elasticity
Gross Margin
–1.5
67%
–2
50%
–3
33%
–4
25%
Thus, if a firm’s gross margin is 50%, its price will be optimal only if its elasticity at that
price is –2. By contrast, if the firm’s elasticity is –3 at its current price, then its pricing
will be optimal only if it yields a gross margin of 33%.
This relationship between gross margin and price elasticity at the optimal price is one
of the principal reasons that marketers take such a keen interest in the price elasticity
Chapter 7 Pricing Strategy
253
of demand. Price elasticities can be difficult to measure, but margins generally are not.
Marketers might now ask whether their current margins are consistent with estimates
of price elasticity. In the next section, we will explore this issue in greater detail.
In the interim, if elasticity changes with price, marketers can use this optimality condition to solve for the optimal price. This condition applies to linear demand functions as
well. Because the optimal price formula for linear demand is relatively simple, however,
marketers rarely use the general optimality condition in this instance.
Data Sources, Complications, and Cautions
The shortcuts for determining optimal prices from linear and constant elasticity
demand functions rest on an assumption that variable costs hold constant over the
range of volumes considered. If this assumption is not valid, marketers will likely find
that a spreadsheet model will offer the easiest way to determine optimal price.
We have explored these relationships in detail because they offer useful perspectives
on the relationship between margins and the price elasticity of demand. In day-to-day
management, margins constitute a starting point for many analyses, including those of
price. One example of this dynamic would be cost-plus pricing.
Cost-plus pricing has received bad press in the marketing literature. It is portrayed not
only as internally oriented, but also as naïve, in that it may sacrifice profits. From an
alternate perspective, however, cost-plus pricing can be viewed as an attempt to maintain
margins. If managers select the correct margin—one that relates to the price elasticity of
demand—then pricing to maintain it may in fact be optimal if demand has constant elasticity. Thus, cost-plus pricing can be more customer-oriented than is widely perceived.
Related Metrics and Concepts
Price Tailoring—a.k.a. Price Discrimination: Marketers have invented a variety of
price discrimination tools, including coupons, rebates, and discounts, for example.
All are designed to exploit variations in price sensitivity among customers. Whenever
customers have different sensitivities to price, or different costs to serve, the astute
marketer can find an opportunity to claim incremental value through price tailoring.
EXAMPLE: The demand for a particular brand of sunglasses is composed of two segments:
style-focused consumers who are less sensitive to price (more inelastic), and value-focused consumers who are more sensitive to price (more elastic) (see Figure 7.8). The style-focused group
has a maximum reservation price of $30 and a maximum willing to buy of 10 units. The valuefocused group has a maximum reservation price of $10 and a maximum willing to buy of 40 units.
254
MARKETING METRICS
Style Segment
60
Demand
50
40
30
20
10
$-
$5
$10
$15
$20
$25
$30
Price
Value Segment
45
40
35
Demand
30
25
20
15
10
5
$-
$5
$10
$15
$20
$25
$30
Price
Figure 7.8 Two Segments Form Demand
Chapter 7 Pricing Strategy
255
Alternative A: One Price for Both Segments
Suppose the sunglasses manufacturer plans to offer one price to both segments. Table 7.11
shows the contribution of several candidate prices. The optimal single price (to the nearest cent) is
$6.77, generating a total contribution of $98.56.
Table 7.11
Two Segments: One Price for Both Segments
Single Price
Value Quantity
Demanded
Style Quantity
Demanded
Total Demand
Total
Contribution
$5
20
8.33
28.33
$85.00
$6
16
8.00
24.00
$96.00
$6.77
12.92
7.74
20.66
$98.56
$7
12
7.67
19.67
$98.33
$8
8
7.33
15.33
$92.00
Alternative B: Price per Segment
If the manufacturer can find a way to charge each segment its own optimal price, it will increase
total contribution. In Table 7.12, we show the optimal prices, quantities, and contributions attainable if each segment pays a distinct optimal price.
Table 7.12
Two Segments: Price Tailoring
MRP
Variable
Costs
Optimal Price
Quantity
Revenue
Contribution
Style
$30
$2
$16
4.67
$74.67
$65.33
Value
$10
$2
$6
16
$96.00
$64.00
20.67
$170.67
$129.33
Total
These optimal prices were calculated as the midpoints between maximum reservation price
(MRP) and variable cost (VC). Optimal contributions were calculated with the formula
Contribution* ⫽ (MWB/MRP) * (P* ⫺ VC)2
In the style-focused segment, for example, this yields
Contribution* = (10/30) * ($16 – $2)2
= 1/3 * 142 = $65.33
256
MARKETING METRICS
Thus, through price tailoring, the sunglasses manufacturer can increase total contribution from
$98.56 to $129.33 while holding quantity constant.
Where variable costs differ between segments, as in an airline’s costs of service in
business class versus economy class, the fundamental calculations are the same. To
determine optimal prices, marketers need only change the variable cost per unit in each
segment to correspond to actual costs.
Caution: Regulation
In most industrial economies, governments have passed regulations concerning price
discrimination. In the United States, the most important of these is the RobinsonPatman Act. According to Supreme Court interpretations of this statute (as of mid2015), Robinson-Patman forbids price discrimination only to the extent that it threatens
to injure competition. Price discrimination to consumers is generally considered legal,
except when implemented on the basis of race, religion, nationality, or gender. There
are two main types of injury contemplated by the Act:
1. Primary line competitive injury: Price discrimination might be used as a predatory tactic. That is, a firm might set prices below cost to certain customers in
order to harm competition at the supplier level. Anti-trust authorities apply this
standard to predatory pricing claims under the Sherman Act and the Federal
Trade Commission Act in order to evaluate allegations of price discrimination.
2. Secondary line competitive injury: A seller that charges different prices to
competing buyers of the same commodity, or that discriminates in providing
“allowances”—such as compensation for advertising or other services—may be
violating the Robinson-Patman Act. Such discrimination can impair competition by awarding favored customers an edge that has nothing to do with superior efficiency.
In the United States, price discrimination is often lawful, particularly if it reflects different costs of dealing with diverse buyers, or if it results from a seller’s attempts to meet
a competitor’s prices or services.6 The situation isn’t always perfectly clear as relevant
laws can differ between states and in some states the courts are yet to rule on pertinent
cases. Clearly, this is not intended to be a legal opinion, however. Legal advice should
be sought for a company’s individual circumstances.
Chapter 7 Pricing Strategy
257
7.5 “Own,” “Cross,” and “Residual” Price Elasticity
The concept of residual price elasticity introduces competitive dynamics into the
pricing process. It incorporates competitor reactions and cross elasticity. This, in
turn, helps explain why prices in daily life are rarely set at the optimal level suggested
by a simpler view of elasticity. Marketers consciously or unconsciously factor competitive dynamics into their pricing decisions.
Residual Price Elasticity (I) = Own Price Elasticity (I) + Competitor Reaction
Elasticity (I) * Cross Elasticity (I)
The greater the competitive reaction anticipated, the more residual price elasticity
will differ from a company’s own price elasticity.
Purpose: To account for both customers’ price elasticity and potential competitive reactions when planning price changes.
Often, in daily life, price elasticity doesn’t quite correspond to the relationships discussed in the prior section. Managers may find, for example, that their estimates of this
key metric are not equal to the negative of the reciprocal of their margins. Does this
mean they’re setting prices that are not optimal? Perhaps.
It is more likely, however, that they’re including competitive factors in their pricing
decisions. Rather than using elasticity as estimated from current market conditions,
marketers may estimate—or intuit—what elasticity will be after competitors respond
to a proposed change in price. This introduces a new concept, residual price elasticity—
customers’ elasticity of demand in response to a change in price, after accounting for
any increase or decrease in competitors’ prices that may be triggered by the initial
change.
Residual price elasticity is the combination of three factors:
1. “Own” price elasticity—The change in units sold due to the reaction of a
firm’s customers to its changes in price.
2. “Competitor reaction” elasticity—The reaction of competitors to a firm’s price
changes.
3. “Cross” price elasticity—The reaction of a firm’s customers to price changes
by its competitors.
These factors and their interactions are illustrated in Figure 7.9.
Own Price Elasticity: How customers in the market react to our price changes.
258
MARKETING METRICS
Competitive Reaction Elasticity: How our competitors respond to our price
changes.
Cross Elasticity: How our customers respond to the price changes of our
competitors.
Our Price
Change
E2: Competitor
Reaction Elasticity
E1: Own Price
Elasticity
Competitor
Price Change
Our Volume
Change
E3: Cross Elasticity
E1 = Own Price Elasticity
E2 = Competitor Reaction Elasticity
E3 = Cross Elasticity
E1 + E2*E3 = Residual Elasticity
Figure 7.9 Residual Price Elasticity
The distinction between own and residual price elasticity is not made clear in the literature. Some measures of price elasticity, for example, incorporate past competitive reactions and thus are more indicative of residual price elasticity. Others principally reflect
own price elasticity and require further analysis to determine where sales and income
will ultimately settle. The following sequence of actions and reactions is illustrative:
1. A firm changes price and observes the resulting change in sales. As an alternative, it may track another measure correlated with sales, such as share of choice
or preference.
2. Competitors observe the firm’s change in price and its increase in sales, and/or
their own decrease in sales.
Chapter 7 Pricing Strategy
259
3. Competitors decide whether and by how much to change their own prices. The
market impact of these changes will depend on (1) the direction and degree of
the changes, and (2) the degree of cross elasticity, that is, the sensitivity of the
initial firm’s sales quantity to changes in competitors’ prices. Thus, after tracking the response to its own price change, the initial firm may observe a further
shift in sales as competitors’ price changes take effect in the market.
Due to this dynamic, if a firm measures price elasticity only through customer response
to its initial actions, it will miss an important potential factor: competitive reactions
and their effects on sales. Only monopolists can make pricing decisions without regard
to competitive response. Other firms may neglect or decline to consider competitive
reactions, dismissing such analyses as speculation. But this generates a risk of shortsightedness and can lead to dangerous surprises. Still other firms may embrace game
theory and seek a Nash Equilibrium to anticipate where prices will ultimately settle. (In
this context, the Nash Equilibrium would be the point at which none of the competitors
in a market have a profit-related incentive to change prices.)
Although a detailed exploration of competitive dynamics is beyond the scope of this
book, we offer a simple framework for residual price elasticity next.
Construction
To calculate residual price elasticity, three inputs are needed:
1. Own price elasticity: The change in a firm’s unit sales, resulting from its initial
price change, assuming that competitors’ prices remain unchanged.
2. Competitor reaction elasticity: The extent and direction of the price changes
that are likely to be made by competitors in response to a firm’s initial price
change. If competitor reaction elasticity is 0.5, for example, then as a firm reduces
its prices by a small percentage, competitors can be expected to reduce their own
prices by half that percentage. If competitor reaction elasticity is –0.5, then as
a firm reduces its prices by a small percentage, competitors will increase their
prices by half that percentage. This is a less common scenario, but it is possible.
3. Cross elasticity with regard to competitor price changes: The percentage
and direction of the change in the initial firm’s sales that will result from a
small percentage change in competitors’ prices. If cross elasticity is 0.25, then
a small percentage increase in competitors’ prices will result in an increase
of one-fourth that percentage in the initial firm’s sales. Note that the sign of
cross elasticity is generally the reverse of the sign of own price elasticity. When
competitors’ prices rise, a firm’s sales will usually increase, and vice versa.
Residual Price Elasticity (I) = Own Price Elasticity (I) + Competitor Reaction
Elasticity (I) * Cross Elasticity (I)
260
MARKETING METRICS
The percentage change in a firm’s sales can be approximated by multiplying its own
price change by its residual price elasticity:
Change in Sales from Residual Elasticity (%) = Own Price Change (%) * Residual Price
Elasticity (I)
Forecasts of any change in sales to be generated by a price change thus should take into
account the subsequent competitive price reactions that can be reasonably expected, as
well as the second-order effects of those reactions on the sales of the firm making the
initial change. The net effect of adjusting for such reactions might be to amplify, diminish, or even reverse the direction of the change in sales that was expected from the initial
price change.
EXAMPLE: A company decides to reduce price by 10% (price change = –10%). It has estimated
its own price elasticity to be –2. Ignoring competitive response, the company would expect a 10%
price reduction to yield an approximately 20% increase in sales (–2 * –10%). (Note: As observed in
our earlier discussion of elasticity, projections based on point elasticity are accurate only for linear
demand functions. Because this example does not specify the shape of the demand function, the
projected 20% increase in sales is an approximation.)
The company estimates competitor reaction elasticity to be 1. That is, in response to the firm’s
action, competitors are expected to shift pricing in the same direction and by an equal percentage.
The company estimates cross elasticity to be 0.7. That is, a small percentage change in competitors’ prices will result in a change in the firm’s own sales of 0.7 percent. On this basis,
Residual Elasticity = Own Price Elasticity + Competitor Reaction Elasticity
* Cross Elasticity
= –2 + 1 * 0.7
= –2 + 0.7
= –1.3
Sales Increase ≈ Change in Price * Residual Elasticity
= –10% * –1.3
= 13% Increase in Sales
Competitor reactions and cross elasticity are expected to reduce the firm’s initially projected
sales increase from 20% to 13%.
Chapter 7 Pricing Strategy
261
Data Sources, Complications, and Cautions
Accounting for potential competitive reactions is important, but there may be simpler
and more reliable methods of managing price strategy in a contested market. Game
theory and price leadership principles offer some guidance.
It is important for managers to distinguish between price elasticity measures that are
inherently unable to account for competitive reactions and those that may already
incorporate some competitive dynamics. For example, in “laboratory” investigations
of price sensitivity—such as surveys, simulated test markets, and conjoint analyses—
consumers may be presented with hypothetical pricing scenarios. These can measure
both own price elasticity and the cross elasticities that result from specific combinations
of prices. But an effective test is difficult to achieve.
Econometric analysis of historical data, evaluating the sales and prices of firms in a
market over longer periods of time (that is, annual or quarterly data), may be better
able to incorporate competitive changes and cross elasticities. To the extent that a firm
has changed price somewhat randomly in the past, and to the extent that competitors
have reacted, the estimates of elasticity that are generated by such analyses will measure
residual elasticity. Still, the challenges and complexities involved in measuring price elasticity from historical data are daunting.
By contrast, short-term test market experiments are unlikely to yield good estimates
of residual price elasticity. Over short periods, competitors might not learn of price
changes or have time to react. Consequently, elasticity estimates based on test markets
are much closer to own price elasticity.
Less obvious, perhaps, are econometric analyses based on transactional data, such as
scanner sales and short-term price promotions. In these studies, prices decline for a
short time, rise again for a longer period, decline briefly, rise again, and so forth. Even if
competitors conduct their own price promotions during the study period, estimates of
price elasticity derived in this way are likely to be affected by two factors. First, competitors’ reactions likely will not be factored into an elasticity estimate because they won’t
have had time to react to the initial firm’s pricing moves. That is, their actions will have
been largely motivated by their own plans. Second, to the extent that consumers stock
up during price deals, any estimates of price elasticity will be higher than would be
observed over the course of long-term price changes.
Prisoner’s Dilemma Pricing
Prisoner’s dilemma pricing describes a situation in which the pursuit of self-interest by
all parties leads to sub-optimal outcomes for all. This phenomenon can lead to stability
at prices above the expected optimal price. In many ways, these higher-than-optimal
prices have the appearance of cartel pricing. But they can be achieved without explicit
262
MARKETING METRICS
collusion, provided that all parties understand the dynamics, as well as their competitors’ motivations and economics.
The prisoner’s dilemma phenomenon derives its name from a story illustrating the
concept. Two members of a criminal gang are arrested and imprisoned. Each prisoner
is placed in solitary confinement, with no means of speaking to the other. Because the
police don’t have enough evidence to convict the pair on the principal charge, they plan
to sentence both to a year in prison on a lesser charge. First, however, they try to get
one or both to confess. Simultaneously, they offer each prisoner a Faustian bargain. If
the prisoner testifies against his partner, he will go free, while the partner is sentenced
to three years in prison on the main charge. But there’s a catch . . . If both prisoners
testify against each other, both will be sentenced to two years in jail.7 On this basis, each
prisoner reasons that he’ll do best by testifying against his partner, regardless of what
the partner does.
For a summary of the choices and outcomes in this dilemma, please see Figure 7.10,
which is drawn in the first person from the perspective of one of the prisoners. Firstperson outcomes are listed in bold. Partner outcomes are italicized.
My
partner
refuses
to
testify
3 years
My
partner
testifies
2 years
1 year
I go free
My partner
goes free
2 years
I testify
1 year
3 years
I refuse to testify
Figure 7.10 Prisoner’s Dilemma Pay-off Grid
Continuing the first-person perspective, each prisoner reasons as follows: If my partner testifies, I’ll be sentenced to two years in prison if I testify as well, or three years if
I don’t. On the other hand, if my partner refuses to testify, I’ll go free if I testify, but
serve one year in prison if I don’t. In either case, I do better if I testify. But this raises a
dilemma. If I follow this logic and testify—and my partner does the same—we end up
in the lower-left cell of the table, serving two years in prison.
Figure 7.11 uses arrows to track these preferences—a dark arrow for the first-person
narrator in this reasoning, and a light arrow for his partner.
Chapter 7 Pricing Strategy
263
My
Partner
Refuses
1 year
3 years
I go free
My
Partner
Testifies
1 year
My partner
goes free
2 years
2 years
I testify
3 years
I refuse
Figure 7.11 Pay-off Grid with Arrows Representing Preferences for Prisoners
The dilemma, of course, is that it seems perfectly logical to follow the arrows and testify. But when both prisoners do so, they both end up worse off than they would have
if they’d both refused. That is, when both testify, both are sentenced to two years in
prison. If both had refused, they both could have shortened that term to a single year.
Admittedly, it takes a good deal of time to grasp the mechanics of the prisoner’s dilemma,
and far longer to appreciate its implications. But the story serves as a powerful metaphor,
encapsulating a wide range of situations in which acting in one’s own best interest leads
to outcomes in which everyone is worse off.
In pricing, there are many situations in which a firm and its competitors face a prisoner’s dilemma. Often, one firm perceives that it could increase profits by reducing prices,
regardless of competitors’ pricing policies. Simultaneously, its competitors perceive
the same forces at work. That is, they too could earn more by cutting prices, regardless
of the initial firm’s actions. If both the initial firm and its competitors reduce prices,
however—that is, if all parties follow their own unilateral best interests—they will, in
many situations, all end up worse off. The industry challenge in these situations is to
keep prices high despite the fact that each firm will benefit by lowering them.
Given a choice between high and low prices a firm faces a prisoner’s dilemma pricing
situation when the following conditions apply:
1. Its contribution is greater at the low price when selling against both high and
low competitor prices.
2. Competitors’ contribution is greater at their low price when selling against both
the high and low prices of the initial firm.
264
MARKETING METRICS
3. For both the initial firm and its competitors, however, contribution is lower
if all parties set their price low than it would have been if all parties had
priced high.
EXAMPLE: As shown in Table 7.13, my firm faces one main competitor. Currently my price is
$2.90, their price is $2.80, and I hold a 40% share of a market that totals 20 million units. If I reduce
my price to $2.60, I expect my share will rise to 55%—unless, of course, they also cut their price. If
they also reduce price by $0.30—to $2.50—then I expect our market shares to remain constant at
40/60. On the other hand, if my competitor cuts its price but I hold steady at $2.90, then I expect
they’ll increase their market share to 80%, leaving me with only 20%.
If we both have variable costs of $1.20 per unit, and market size remains constant at 20 million
units, we face four possible scenarios with eight contribution figures—four for my firm and four for
the competition:
Table 7.13
Scenario Planning Pay-off Table
My Volume
(m)
My Sales
($m)
My
My Variable Contribution
Costs ($m) ($m)
My Firm High. $2.90
Competition
High.
8
$23.2
$9.6
$13.6
My Firm High. $2.90
Competition
Low.
4
$11.6
$4.8
$6.8
My Firm Low.
Competition
Low.
$2.60
8
$20.8
$9.6
$11.2
My Firm Low.
Competition
High.
$2.60
11
$28.6
$13.2
$15.4
Pricing
Scenario
My Price
Chapter 7 Pricing Strategy
265
Their Volume Their Sales
(m)
($m)
Their
Variable
Costs ($m)
Their
Contribution
($m)
My Firm High. $2.80
Competition
High.
12
$33.6
$14.4
$19.2
My Firm High. $2.50
Competition
Low.
16
$40.0
$19.2
$20.8
My Firm Low.
Competition
Low.
$2.50
12
$30.0
$14.4
$15.6
My Firm Low.
Competition
High.
$2.80
9
$25.2
$10.8
$14.4
Pricing
Scenario
Their Price
Are we in a prisoner’s dilemma situation?
Figure 7.12 shows the four contribution possibilities for both my firm and my competitor.
Their
Price
= $2.80
$14.4
$19.2
$15.4
$13.6
High
Their
Price
= $2.50
$20.8
$15.6
$11.2
Low
My Price = $2.60
Low
Figure 7.12
$6.8
My Price = $2.90
High
Pay-off Grid with Expected Values (Values Are in the Millions of Dollars)
Let’s check to see whether the conditions for the prisoner’s dilemma are met:
1. My contribution is higher at the low price for both high and low competitor prices
($15.4m > $13.6m, and $11.2m > $6.8m). No matter what my competitor does, I
make more money at the low price.
266
MARKETING METRICS
2. My competitor’s contribution is higher at the low price, regardless of my price
($15.6m > $14.4m, and $20.8m > $19.2m). They, too, are better off at the low price,
regardless of my price.
3. For both my firm and my competitor, however, contribution is lower if we both
price low than it would be if we both price high ($15.6m < $19.2m, and $11.2m <
$13.6m).
The conditions for the prisoner’s dilemma are met (see Figure 7.13).
Their
Price
= $2.80
$14.4
$19.2
$15.4
Their
Price
= $2.50
$13.6
$20.8
$15.6
$11.2
My Price = $2.60
$6.8
My Price = $2.90
Figure 7.13 Pay-off Grid with Expected Values and Preference Arrows
(Contribution Values Are in the Millions of Dollars)
The implication for my firm is clear: Although it is tempting to lower my price, seeking increased
share and a $15.4 million contribution, I must recognize that my competitor faces the same incentives. They, too, have an incentive to cut price, grab share, and increase their contribution. But if
they lower their price, I’ll probably lower mine. If I lower my price, they’ll probably lower theirs. If
we both reduce our prices, I’ll earn only $11.2m in contribution—a sharp decline from the $13.6m
I make now.
Managerial Note: To determine whether you face a prisoner’s dilemma situation,
project the dollar contributions for both your firm and your competition at four combinations of high and low prices. Projections may require assumptions about your competitors’ economics. These, in turn, will require care. If competitors’ economics differ
greatly from your projections, they may not face the decisions or motivations ascribed
to them in your model. Additionally, there are a number of reasons why the logic of the
prisoner’s dilemma won’t always hold, even if all assumptions are correct.
Chapter 7 Pricing Strategy
267
1. Contribution may not be the sole criterion in decision-making: In our
example, we used contribution as the objective for both firms. Market share,
however, may have importance to one or more firms, above and beyond its
immediate, direct effect on contribution. Whatever a firm’s objective may be, if
it is quantifiable, we can place it in our table to better understand the competitive situation.
2. Legal issues: Certain activities designed to discourage competition and maintain high prices are illegal. Our purpose here is to help managers understand
the economic trade-offs involved in competitive pricing. Managers should be
aware of their legal environment and behave accordingly.
3. Multiple competitors: Pricing becomes more complicated when there are multiple competitors. The test for a multi-party prisoner’s dilemma is the logical
extension of the test described earlier. A major difference, however, arises in
practice. As a general principle, the greater the number of independent competitors, the more difficult it will be to keep prices high.
4. Single versus repeated play: In our original story, two prisoners decide whether to testify in a single investigation. In game theory terms, they play the game a
single time. Experiments have shown that in a single play of a prisoner’s dilemma, the likely outcome is that both prisoners will testify. If the game is played
repeatedly, however, it is more likely that both prisoners will refuse to testify.
Because pricing decisions are made repeatedly, this evidence suggests that high
prices are a more likely outcome. Most businesses eventually learn to live with
their competition.
5. More than two possible prices: We have examined a situation in which each
player considers two prices. In reality, there may be a wide range of prices
under consideration. In such situations, we might extend our analysis to more
boxes. Once again, we might add arrows to track preferences. Using these more
complex views, one sometimes finds areas within the table in which a prisoner’s dilemma applies (usually at the higher prices), and others where it does
not (usually at the lower prices). One might also find that the arrows lead to
a particular cell in the middle of the table called the equilibrium. A prisoner’s
dilemma situation generally applies for prices higher than the set of equilibrium
prices.
Applying the lessons of the prisoner’s dilemma, we see that optimal price calculations
based on own price elasticity may lead us to act in our own unilateral best interest.
By contrast, when we factor residual price elasticity into our calculations, competitive
response becomes a key element of our pricing strategy. As the prisoner’s dilemma
shows, over the long term, a firm is not always best served by acting in its apparent
unilateral best interest.
268
MARKETING METRICS
References and Suggested Further Reading
Dolan, Robert J., and Hermann Simon. (1996). Power Pricing: How Managing Price
Transforms the Bottom Line, New York: Free Press, 4.
Roegner, E.V., M.V. Marn, and C.C. Zawada. (2005). “Pricing,” Marketing Management,
14(1), 23–28.
Chapter 7 Pricing Strategy
269
This page intentionally left blank
8
PROMOTION
Introduction
Key concepts covered in this chapter:
Baseline Sales, Incremental Sales,
and Promotional Lift
Percent Sales on Deal, Percent Time
on Deal, and Average Deal Depth
Redemption Rates for Coupons/
Rebates
Pass-Through and Price
Waterfall
Price promotions can be divided into two broad categories:
■
Temporary price reductions.
■
Permanent features of pricing systems.1
With both of these, firms seek to change the behavior of consumers and trade customers in ways that increase sales and profits over time, though a promotion’s short-term
effect on profits will often be negative. There are multiple routes to sales and profit
growth and many potential reasons for offering price promotions. Such programs
might be aimed at affecting the behavior of end users (consumers), trade customers
(distributors or retailers), competitors, or even a firm’s own salespeople. Although
the goal of a promotion is often to increase sales, these programs can also affect costs.
Examples of specific, short-term promotional objectives include the following:
■
■
To acquire new customers, perhaps by generating trial.
To appeal to new or different segments that are more price-sensitive than a
firm’s traditional customers.
271
■
To increase the purchase rates of existing customers; to increase loyalty.
■
To gain new trade accounts (that is, distribution).
■
To introduce new SKUs to the trade.
■
To increase shelf space.
■
■
To blunt competitive efforts by encouraging the firm’s customers to “load up”
on inventory.
To smooth production in seasonal categories by inducing customers to order
earlier (or later) than they ordinarily would.
The metrics for many of these interim objectives, including trial rate and percentage
of new product sales, are covered elsewhere. In this chapter, we focus on metrics for
monitoring the acceptance of price promotions and their effects on sales and profits.
The most powerful framework for evaluating temporary price promotions is to partition sales into two categories: baseline and incremental. Baseline sales are those that a
firm would have expected to achieve if no promotion had been run. Incremental sales
represent the “lift” in sales resulting from a price promotion. By separating baseline
sales from incremental lift, managers can evaluate whether the sales increase generated
by a temporary price reduction compensates for the concomitant decrease in prices
and margins. Similar techniques are used in determining the profitability of coupons
and rebates.
Although the short-term effect of a price promotion is almost invariably measured by
its increase in sales, over longer periods management becomes concerned about the
percentage of sales on deal and the percentage of time during which a product is on
deal. In some industries, list price has become such a fiction that it is used only as a
benchmark for discussing discounts.
Average deal depth and the price waterfall help capture the depth of price cuts and
explain how one arrives at a product’s net price (pocket price) after accounting for all
discounts. There are often major differences between the discounts offered to trade
customers and the extent to which those discounts are accepted. There may also be a
difference between the discounts received by the trade and those that the trade shares
with its customers. The pass-through percentage and price waterfall are analytic
structures designed to capture those dynamics and thus to measure the impact of a
firm’s promotions.
272
MARKETING METRICS
Metric
Construction
Considerations
Purpose
8.1
Baseline Sales
Intercept in
Marketing activities
regression of
also contribute to
sales as function
baseline.
of marketing
variables. Baseline
Sales = Total Sales,
less incremental
sales generated by a
marketing program
or programs.
8.1
Incremental Sales, Total sales, less
Need to consider
or Promotional Lift baseline sales.
competitive
Regression
actions.
coefficient to
marketing variables
cited above.
To determine
short-term effects
of marketing effort.
8.2
Redemption Rates
Coupons redeemed Will differ
divided by coupons significantly by
distributed.
mode of coupon
distribution.
Rough measure of
coupon “lift” after
adjusting for sales
that would have
been made without
coupons.
8.2
Costs for Coupons
and Rebates
Coupon face
amount plus
redemption
charges, multiplied
by the number of
coupons redeemed.
Does not consider
margins that
would have been
generated by those
willing to buy
product without
coupon.
Allows for
budgeting of
coupon expense.
8.2
Percentage Sales
with Coupon
Sales via coupon,
divided by total
sales.
Doesn’t factor
in magnitude of
discount offered by
specific coupons.
A measure of
brand dependence
on promotional
efforts.
To determine the
extent to which
current sales are
independent of
specific marketing
efforts.
Continues
Chapter 8 Promotion
273
Metric
Construction
Considerations
Purpose
8.3
Percent Sales on
Deal
Sales with
temporary
discounts as a
percentage of total
sales.
Does not make
distinction for
depth of discounts
offered.
A measure of
brand dependence
on promotional
efforts.
8.3
Pass-Through
Promotional
discounts provided
by the trade
to consumers,
divided by
discounts provided
to the trade by the
manufacturer.
Can reflect power
in the channel,
or deliberate
management or
segmentation.
To measure the
extent to which
a manufacturer’s
promotions
generate
promotional
activity
further along
the distribution
channel.
8.4
Price Waterfall
Actual average
price per unit
divided by list price
per unit. Can also
be calculated by
working backward
from list price,
taking account of
potential discounts,
weighted by the
frequency with
which each is
exercised.
Some discounts
may be offered at
an absolute level,
not on a per-item
basis.
To indicate the
price actually paid
for a product, and
the sequence of
channel factors
affecting that price.
274
MARKETING METRICS
8.1 Baseline Sales, Incremental Sales,
and Promotional Lift
Estimates of baseline sales establish a benchmark for evaluating the incremental sales
generated by specific marketing activities. This baseline also helps isolate incremental sales from the effects of other influences, such as seasonality or competitive promotions. The following equations can be applied for defined periods of time and for
the specific element of the marketing mix that is used to generate incremental sales.
Total Sales ($,#) = Baseline Sales ($,#) + Incremental Sales from Marketing ($,#)
Incremental Sales from Marketing ($,#) = Incremental Sales from Advertising ($,#)
+ Incremental Sales from Trade
Promotion ($,#)
+ Incremental Sales from Consumer
Promotion ($,#)
+ Incremental Sales from Other ($,#)
Lift (from Promotion) (%) =
Cost of Incremental Sales ($) =
Incremental Sales ($,#)
Baseline Sales ($,#)
Marketing Spending ($)
Incremental Sales ($,#)
The justification of marketing spending almost always involves estimating the incremental effects of the program under evaluation. However, because some marketing
costs are often assumed to be fixed (for example, marketing staff and sales force
salaries), one rarely sees incremental sales attributed to these elements of the mix.
Purpose: To select a baseline of sales against which the incremental sales and
profits generated by marketing activity can be assessed.
A common problem in marketing is estimating the sales “lift” attributable to a specific
campaign or set of marketing activities. Evaluating lift entails making a comparison
with baseline sales, the level of sales that would have been achieved without the program under evaluation. Ideally, experiments or “control” groups would be used to
establish baselines. If it were quick, easy, and inexpensive to conduct such experiments,
this approach would dominate. In lieu of such control groups, marketers often use historical sales adjusted for expected growth, taking care to control for seasonal influences.
Regression models that attempt to control for the influence of these other changes
are often used to improve estimates of baseline sales. Ideally, both controllable and
Chapter 8 Promotion
275
uncontrollable factors, such as competitive spending, should be included in baseline
sales regression models. When regression is used, the intercept is often considered to
be the baseline.
Construction
In theory, determining incremental sales is as simple as subtracting baseline sales from
total sales. Challenges arise, however, in determining baseline sales.
Baseline Sales: Expected sales results, excluding the marketing programs under
evaluation.
In reviewing historical data, total sales are known. The analyst’s task then is to
separate these into baseline sales and incremental sales. This is typically done with
regression analysis. The process can also involve test market results and other market
research data.
Total Sales ($,#) = Baseline Sales ($,#) + Incremental Sales ($,#)
Analysts also commonly separate incremental sales into portions attributable to the
various marketing activities used to generate them.
Incremental Sales ($,#) = Incremental Sales from Advertising ($,#) + Incremental Sales
from Trade Promotion ($,#) + Incremental Sales from Consumer
Promotion ($,#) + Incremental Sales from Other ($,#)
Baseline sales are generally estimated through analyses of historical data. Firms often
develop sophisticated models for this purpose, including variables to adjust for market
growth, competitive activity, and seasonality, for example. That done, a firm can use its
model to make forward-looking projections of baseline sales and use these to estimate
incremental sales.
Incremental sales can be calculated as total sales, less baseline sales, for any period of
time (for example, a year, a quarter, or the term of a promotion). The lift achieved by
a marketing program measures incremental sales as a percentage of baseline sales. The
cost of incremental sales can be expressed as a cost per incremental sales dollar or a cost
per incremental sales unit (for example, cost per incremental case).
Incremental Sales ($,#) = Total Sales ($,#) – Baseline Sales ($,#)
Lift (%) =
Cost of Incremental Sales ($) =
276
MARKETING METRICS
Incremental Sales ($,#)
Baseline Sales ($,#)
Marketing Spending ($)
Incremental Sales ($,#)
EXAMPLE: A retailer expects to sell $24,000 worth of light bulbs in a typical month without advertising. In May, while running a newspaper ad campaign that cost $1,500, the store sells
$30,000 worth of light bulbs. It engages in no other promotions or non-recurring events during the
month. Its owner calculates incremental sales generated by the ad campaign as follows:
Incremental Sales ($) = Total Sales ($) – Baseline Sales ($)
= $30,000 – $24,000 = $6,000
The store owner estimates incremental sales to be $6,000. This represents a lift (%) of 25%, calculated as follows:
Lift (%) =
=
Incremental Sales ($)
Baseline Sales ($)
$6,000
$24,000
= 25%.
The cost per incremental sales is $0.25, calculated as follows:
Cost of Incremental Sales ($) =
=
Marketing Spending ($)
Incremental Sales ($)
$1,500
$6,000
= 0.25
Total sales can be analyzed or projected as a function of baseline sales and lift. When
estimating combined marketing mix effects, one must be sure to determine whether
lift is estimated through a multiplicative or an additive equation. Additive equations
combine marketing mix effects as follows:
Total Sales ($,#) = Baseline Sales + [Baseline Sales ($,#) * Lift (%) from Advertising]
+ [Baseline Sales ($,#) * Lift (%) from Trade Promotion]
+ [Baseline Sales ($,#) * Lift (%) from Consumer Promotion]
+ [Baseline Sales ($,#) * Lift (%) from Other]
This additive approach is consistent with the conception of total incremental sales as a
sum of the incremental sales generated by various elements of the marketing mix. It is
equivalent to a statement that
Total Sales ($,#) = Baseline Sales + Incremental Sales from Advertising + Incremental Sales
from Trade Promotion + Incremental Sales from Consumer Promotion
+ Incremental Sales from Other
Chapter 8 Promotion
277
Multiplicative equations, by contrast, combine marketing mix effects by using a multiplication procedure, as follows:
Total Sales ($,#) = Baseline Sales ($,#) * (1 + Lift (%) from Advertising) * (1 + Lift (%) from
Trade Promotion) * (1 + Lift (%) from Consumer Promotion) * (1 + Lift
(%) from Other)
When using multiplicative equations, it makes little sense to talk about the incremental
sales from a single mix element. In practice, however, one may encounter statements
that attempt to do exactly that.
EXAMPLE: Company A collects data from past promotions and estimates the lift it achieves
through different elements of the marketing mix. One researcher believes that an additive model
would best capture these effects. A second researcher believes that a multiplicative model might
better reveal the ways in which multiple elements of the mix combine to increase sales. The product
manager for the item under study receives the two estimates shown in Table 8.1.
Table 8.1
Expected Returns to Marketing Spending
Additive
Multiplicative
Trade
Advertising Promotion
Spending Lift
Lift
Consumer
Trade
Consumer
Promotion Advertising Promotion Promotion
Lift
Lift
Lift
Lift
$0
0%
0%
0%
1
1
1
$100k
5.5%
10%
16.5%
1.05
1.1
1.15
$200k
12%
24%
36%
1.1
1.2
1.3
Fortunately, both models estimate baseline sales to be $900,000. The product manager wants to
evaluate the following spending plan: advertising ($100,000), trade promotion ($0), and consumer promotion ($200,000). He projects sales using each method as follows:
Additive:
Projected Sales ($) = $900,000 + [$900,000 * 5.5%] + [$900,000 * 0] + [$900,000 * 36%]
= $900,000 + $49,500 + $0 + $324,000
= $1,273,500
278
MARKETING METRICS
Multiplicative:
Projected Sales = Baseline * Advertising Lift * Trade Promotion Lift
* Consumer Promotion Lift
= $900,000 * 1.05 * 1 * 1.3
= $1,228,500
Note: Because these models are constructed differently, they will inevitably yield different results at
most levels. The multiplicative method accounts for a specific form of interactions between marketing variables. The additive method, in its current form, does not account for interactions.
When historic sales have been separated into baseline and incremental components,
it is relatively simple to determine whether a given promotion was profitable during
the period under study. Looking forward, the profitability2 of a proposed marketing
activity can be assessed by comparing projected levels of profitability with and without
the program:
Profitability of a Promotion ($) = Profits Achieved with Promotion ($)
– Estimated Profits without Promotion
(that is, Baseline) ($)
EXAMPLE: Fred, the VP of Marketing, and Jeanne, the VP of Finance, receive estimates that
sales will total 30,000 units after erecting special displays. Because the proposed promotion involves
a considerable investment ($100,000), the CEO asks for an estimate of the incremental profit
associated with the displays. Because this program involves no change in price, contribution per
unit during the promotion is expected to be the same as at other times, $12.00 per unit. Thus, total
contribution during the promotion is expected to be 30,000 * $12, or $360,000. Subtracting the
incremental fixed cost of specialized displays, profits for the period are projected to be $360,000 –
$100,000, or $260,000.
Fred estimates that baseline sales total 15,000 units. On this basis, he calculates that contribution
without the promotion would be $12 * 15,000 = $180,000. Thus, he projects that the special displays can be expected to generate incremental profit of $360,000 – $180,000 – $100,000 = $80,000.
Jeanne argues that she would expect sales of 25,000 units without the promotion, generating baseline contribution of $12 * 25,000 = $300,000. Consequently, if the promotion is implemented, she
anticipates an incremental decline in profits from $300,000 to $260,000. In her view, the promotion’s lift would not be sufficient to cover its incremental fixed costs. Under this promotion, Jeanne
believes that the firm would be spending $100,000 to generate incremental contribution of only
$60,000 (that is, 5,000 units * $12 contribution per unit).
The baseline sales estimate is a crucial factor here.
Chapter 8 Promotion
279
EXAMPLE: A luggage manufacturer faces a difficult decision regarding whether to launch
a new promotion. The firm’s data show a major increase in product sales in November and
December, but its managers are unsure whether this is a permanent trend of higher sales or merely
a blip—a successful period that can’t be expected to continue (see Figure 8.1).
Sales
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Figure 8.1 Monthly Sales Patterns
The firm’s VP of Marketing strongly supports the proposed promotion. He argues that the
increased volume can’t be expected to continue and that the firm’s historic baseline (26,028 units)
should be used as the level of sales that can be anticipated without the promotion. In addition, the
Marketing VP argues that only the variable cost of each sale should be considered. “After all, the
fixed costs will be with us whatever we do,” he says. On this basis, the relevant cost per unit subject
to analysis would be $25.76.
The CEO hires a consultant who has a very different opinion. In the consultant’s view, the
November-December sales increase was more than a blip. The market has grown, she says, and the
strength of the firm’s brand has grown with it. Consequently, a more appropriate estimate of baseline sales would be 48,960 units. The consultant also points out that in the long term, no costs are
fixed. Therefore, for purposes of analysis, fixed costs should be allocated to the cost of the product
because the product must ultimately generate a return after such expenses as factory rent are paid.
On this basis, the full cost of each unit, $34.70, should be used as the cost of incremental sales (see
Table 8.2).
280
MARKETING METRICS
Table 8.2
Baseline Matters When Considering Profitability
Consultant
VP Marketing
Promotion
Baseline
Promotion
Baseline
Price
$41.60
$48.00
$41.60
$48.00
Cost
$34.70
$34.70
$25.76
$25.76
Margin
$6.90
$13.30
$15.84
$22.24
Sales
75,174
48,960
75,174
26,028
Profit
$518,701
$651,168
$1,190,756
$578,863
Profitability of
Promotion
($132,467)
$611,893
The Marketing VP and the consultant make very different projections of the profitability of the
promotion. Once again, the choice of the baseline matters. Also, we can see that establishing a
shared understanding of costs and margins can be critical.
Data Sources, Complications, and Cautions
Finding a baseline estimate of what a company can be expected to sell, “all things being
equal,” is a complex and inexact process. Essentially, the baseline is the level of sales
that can be expected without significant marketing activities. When certain marketing
activities, such as price promotions, have been employed for several periods, it can be
especially difficult to separate “incremental” and “baseline” sales.
In many companies, it is common to measure sales performance against historic data.
In effect, this sets historic sales as the baseline level for analysis of the impact of marketing spending. For example, retailers can evaluate their performance on the basis of
same store sales (to remove differences caused by the addition or removal of outlets).
Further, they can compare each current period to the same period in the prior year,
in order to avoid seasonality biases and to ensure that they measure periods of special
activity (such as sales events) against times of similar activity.
It is also common practice to adjust the profitability of promotions for longer-term
effects. These effects can include a decline in sales levels in periods immediately
following a promotion, as well as higher or lower sales in related product categories that
are associated with a promotion. Adjustments can be negative or positive. Additional
Chapter 8 Promotion
281
long-term effects, such as obtaining trial by new consumers, gaining distribution with
trade customers, and increased consumption rates were discussed briefly in the chapter
introduction.
Long-term Effects of Promotions
Over time, the effects of promotions may be to “ratchet” sales up or down (see Figures
8.2 and 8.3). Under one scenario, in response to one firm’s promotions, competitors may
also increase their promotional activity, and consumers and trade customers in the field
may learn to wait for deals, increasing sales for no one (see the prisoner’s dilemma in
Section 7.5).
Profits
Promotion
Aggressive
Promotion
Competitors
react
More
Aggressive
Promotion
Competitors
react
Competitors
react
Customers
learn: wait
for deals
BASELINE
Time
Figure 8.2 Downward Spiral—Promotional Effectiveness
Profits
Promotion 2
Promotion 3
Promotion 1
Customers
try product
Trade stocks
successful
product
Customers
become loyal
More customers
learn to love the
product
BASELINE
Time
Figure 8.3 Successful Promotion with Long-Term Benefits
282
MARKETING METRICS
Under a different, more heartening scenario, promotions can generate trial for new
products, build trade distribution, and encourage loyalty, thus raising the long-term
level of baseline sales.
8.2 Redemption Rates, Costs for Coupons and Rebates,
Percent Sales with Coupon
Redemption rate is the percentage of distributed coupons or rebates that are used
(redeemed) by consumers.
Coupons Redeemed (#)
Coupon Redemption Rate (%) =
Coupons Distributed (#)
Cost per Redemption ($) = Coupon Face Amount ($) + Redemption Charges ($)
Total Coupon Cost ($) = Cost per Redemption ($) * Coupons Redeemed (#)
+ Coupon Printing and Distribution Cost ($)
Percentage Sales with Coupon (%) =
Sales with Coupon ($)
Sales ($)
The redemption rate is an important metric for marketers assessing the effectiveness
of their coupon distribution strategy. It helps determine whether coupons are reaching the customers who are motivated to use them. Similar metrics apply to mail-in
rebates.
Cost per redemption ($) measures variable costs per coupon redeemed. Coupon
distribution costs are usually viewed as fixed costs.
Purpose: To track and evaluate coupon usage.
Some people hate coupons. Some like them. And some say they hate coupons, but
really like them. Businesses often say they hate coupons but continue to use them.
Coupons and rebates are used to introduce new products, to generate trial of existing
products by new customers, and to “load” consumers’ pantries, encouraging long-term
consumption.
Almost all of the interim objectives discussed in the introduction to this chapter can apply
to coupons and rebates. Coupons can be used to offer lower prices to more price-sensitive consumers. Coupons also serve as a form of advertising, making them dual-purpose
marketing vehicles. Coupon clippers will see a brand name and pay closer attention to
it—considering whether they desire the product—than would an average consumer
exposed to an advertisement without a compelling offer. Finally, both rebates and coupons can serve as focus points for retailer promotions. To generate traffic, retailers can
Chapter 8 Promotion
283
double or even triple coupon amounts—generally up to a declared limit. Retailers also
often advertise prices “after rebates” in order to promote sales and perceptions of value.
Construction
Coupon Redemption Rate (%) =
Coupons Redeemed (#)
Coupons Distributed (#)
Cost per Redemption ($) = Coupon Face Amount ($) + Redemption Charges ($)
Total Coupon Cost: Reflects distribution, printing,3 and redemption costs to estimate the total cost of a coupon promotion.
Total Coupon Cost ($) = Coupons Redeemed (#) * Cost per Redemption ($)
+ Coupon Printing and Distribution Cost ($)
Total Cost per Redemption ($) =
Percentage Sales with Coupon (%) =
Total Coupon Cost ($)
Coupons Redeemed (#)
Sales with Coupon ($,#)
Sales ($,#)
To determine the profitability of coupons and rebates, managers require approaches
similar to those used in estimating baseline and incremental sales, as discussed in the
previous section of this chapter. By themselves, redemption rates are not a good measure
of success. Under certain circumstances, even low redemption rates can be profitable.
Under other circumstances, by contrast, high redemption rates can be quite damaging.
EXAMPLE: Yvette is the Manager of Analysis for a small regional consumer packaged goods
firm. Her product has a dominant share of the retail distribution in a narrow geographic area. Her
firm decides to launch a coupon campaign, and Yvette is charged with reporting on the program’s
success. Her assistant looks at the figures and realizes that of the 100,000 coupons distributed in
the local paper, 5,000 were used to buy product. The assistant is excited when he calculates that this
represents a 5% redemption rate—a much higher figure than the company has ever previously seen.
Yvette, however, is more cautious in judging the promotion a success. She checks the sales of the
relevant product and learns that these increased by only 100 units during the promotion period.
Yvette concludes that the vast majority of coupon use was by customers who would have bought the
product anyway. For most customers, the sole impact of the coupon was to reduce the price of the
product below the level they would have willingly paid. Until she conducts a full profitability analysis, evaluating the profit generated by the 100 incremental sales and comparing this to coupon costs
and the value lost on most coupon sales, Yvette can’t be sure that the program made an overall loss.
But she feels certain that she should curtail the celebrations.
284
MARKETING METRICS
Data Sources, Complications, and Cautions
To calculate coupon redemption rates, managers must know the number of coupons
placed in circulation (distributed) as well as the number redeemed. Companies generally engage distribution services or media companies to place coupons in circulation.
Redemption numbers are usually derived from the invoices presented by coupon
clearinghouses.
Related Metrics and Concepts
Mail-in Rebates
The rebate, in effect, is a form of coupon that is popular with big-ticket items. Its usage
dynamics are straightforward: Customers pay the full price for a product, enabling
retailers to meet a specific price point. The customer then exercises the rebate and
receives back a specified dollar amount.
By using rebates, marketers gain information about customers, which can be useful
in remarketing and product control. Mail-in rebates also reduce the effective price
of an item for customers who are sufficiently price-conscious to take advantage of
them. Others pay full price. The “non-redemption rates” for rebates are sometimes
called “breakage.”
Breakage: The number of rebates not redeemed by customers. The breakage rate
is the percentage of rebates not redeemed.
EXAMPLE: A cell phone company sold 40,000 handsets in one month. On each purchase, the
customer was offered a $30 rebate. Thirty thousand rebates were successfully claimed.
In volume terms, the rebate redemption rate can be calculated by dividing the number of rebates
successfully claimed (30,000) by number offered (40,000):
Redemption Rate (in volume terms) =
30,000
40,000
= 75%
Managers often balk at the cost of distributing coupons. Because promotions rely on
adequate distribution, however, it is inadvisable to create arbitrary cutoffs for distribution costs. The total cost of incremental sales generated would represent a better metric
to evaluate coupon efficiency—and thus to determine the point at which diminishing
returns make further coupon distribution unattractive.
Chapter 8 Promotion
285
In evaluating a coupon or rebate program, companies should also consider the overall
level of benefit provided to consumers. Retailers commonly increase the value of coupons, offering customers a discount of double or even triple the coupons’ face value.
This enables retailers to identify price-sensitive customers and offer them additional
savings. Of course, by multiplying the savings afforded consumers, the practice of doubling or tripling coupons undoubtedly raises some redemption rates.
8.3 Promotions and Pass-Through
Of the promotional value provided by a manufacturer to its retailers and distributors, the pass-through percentage represents the portion that ultimately reaches
the consumer.
Sales with any Temporary Discount ($,#)
Percentage Sales on Deal (%) =
Total Sales ($,#)
Pass-Through (%) =
Value of Temporary Promotional Discounts
Provided to Consumers by the Trade ($)
Value of Temporary Discounts Provided to Trade
by Manufacturer ($)
Manufacturers offer many discounts to their distributors and retailers (often called
“the trade”) with the objective of encouraging them to offer their own promotions,
in turn, to their customers. If trade customers or consumers do not find promotions
attractive, this will be indicated by a decline in percentage sales on deal. Likewise,
low pass-through percentages can indicate that too many deals—or the wrong kinds
of deals—are being offered.
Purpose: To measure whether trade promotions are generating
consumer promotions.
Pass-Through: The percentage of the value of manufacturer promotions paid to
distributors and retailers that is reflected in discounts provided by the trade to
their own customers.
“Middlemen” are a part of the channel structure in many industries. Companies may
face one, two, three, or even four levels of “resellers” before their product reaches the
ultimate consumer. For example, a beer manufacturer may sell to an exporter, who sells
to an importer, who sells to a local distributor, who sells to a retail store. If each channel
adds its own margin, without regard for how others are pricing, the resulting price can
be higher than a marketer would like. This sequential application of individual margins
has been referred to as “double marginalization.”4
286
MARKETING METRICS
Construction
Percentage Sales on Deal: Measures the percentage of company sales that are
sold with a temporary trade discount of some form. Note: This usually would not
include standard discounts such as those for early payment or cooperative advertising allowances (accruals).
Percentage Sales on Deal (%) =
Sales with Any Temporary Discount (#,$)
Total Sales (#,$)
Promotional discount represents the total value of promotional discounts given
throughout the sales channel.
Promotional Discount ($) = Sales with Any Temporary Discount ($)
* Average Depth of Discount As Percent of List (%)
Unit Discount ($)
Depth of Discount As Percent of List =
Unit List Price ($)
Pass-through is calculated as the value of discounts given by the trade to their customers, divided by the value of temporary discounts provided by a manufacturer to the
trade.
Pass-Through (%) =
Promotional Discounts Provided by the Trade to Consumers ($)
Discounts Provided to Trade by Manufacturer ($)
Data Sources, Complications, and Cautions
Manufacturers often compete with one another for the attention of retailers, distributors, and other resellers. Toward that end, they build special displays for their products,
change assortments to include new offerings, and seek to elicit increasing attention
from resellers’ sales personnel. Significantly, in their effort to increase channel “push,”
manufacturers also offer discounts and allowances to the trade. It is important to
understand the rates and amounts of discounts provided to the trade, as well as the
proportions of those discounts that are passed along to the resellers’ customers. At
times, when resellers’ margins are thin, manufacturers’ discounts are designed to
enhance them. Market leaders often worry that trade margins are too thin to support
push efforts. Other manufacturers may be concerned that retail margins are too high,
and that too few of their discounts are being passed along. The metrics discussed in this
chapter should be interpreted with these thoughts in mind.
Resellers may decide that optimizing an entire product line is more important than
maximizing profits on any given product. If a reseller stocks multiple competing lines,
it can be difficult to find an overall solution that suits both that reseller and its suppliers.
Manufacturers strive to motivate resellers to market their goods aggressively and to
grow their shared sales through such programs as incentives for “exclusivity,” or rebates
based on increasing shares of category sales or on year-to-year growth in sales.
Chapter 8 Promotion
287
Resellers learn to adapt their buying and selling practices to take advantage of manufacturer pricing incentives. In this area, marketers must pay special attention to the law of
unforeseen consequences. For example, resellers have been known to
■
■
Buy larger quantities of a product than they can sell—or want to sell—in order
to qualify for volume discounts. The excess goods are then sold (diverted) to
other retailers, stored for future sales, or even destroyed or returned to the
manufacturer for “credit.”
Time their purchases at the ends of accounting periods in order to qualify for
rebates and allowances. This results in “lumpy” sales patterns for manufacturers, making forecasting difficult, increasing problems with out-of-date products
and returns, and raising production costs.
In some instances, a particularly powerful channel “captain” can impose pricing discipline on an entire channel. In most cases, however, each “link” in the distribution
chain can coordinate only its own pricing. A manufacturer, for example, may work out
appropriate pricing incentives for wholesalers, and the wholesalers in turn may develop
their own pricing incentives for retailers.
In many countries and industries, it is illegal for suppliers to dictate the selling prices
of resellers. Manufacturers can’t dictate wholesaler selling prices, and wholesalers can’t
dictate retail prices. Consequently, members of the channel seek indirect methods of
influencing resellers’ prices.
8.4 Price Waterfall
The price waterfall is a way of describing the progression of prices from published
list price to the final price paid by a customer. Each drop in price represents a drop
in the “water level.” For example:
100
List Price
Dealer Discount
90
Cash Discount
85
Annual Rebate
82
Co-op Advertising
Net Price $80
288
MARKETING METRICS
Price Waterfall (%) =
Net Price per Unit ($)
List Price per Unit ($)
In this structure, the average price paid by customers will depend on the list price
of a product, the sizes of discounts given, and the proportion of customers taking
advantage of those discounts.
By analyzing the price waterfall, marketers can determine where product value is
being lost. This can be especially important in businesses that allow the sales channel to reduce prices in order to secure customers. The price waterfall can help focus
attention on deciding whether these discounts make sense for the business.
Purpose: To assess the actual price paid for a product, in comparison with
the list price.
In pricing, the bad news is that marketers can find it difficult to determine the right list
price for a product. The good news is that few customers will actually pay that price
anyway. Indeed, a product’s net price—the price actually paid by customers—often falls
between 53% and 94% of its base price.5
Net Price: The actual price paid for a product by customers after all discounts
and allowances have been factored in. Also called the pocket price.
List Price: The price of a good or service before discounts and allowances are
considered.
Invoice Price: The price specified on the invoice for a product. This price
will typically be stated net of some discounts and allowances, such as dealer,
competitive, and order size discounts, but will not reflect other discounts and
allowances, such as those for special terms and cooperative advertising. Typically,
the invoice price will therefore be less than the list price but greater than the
net price.
Price Waterfall: The reduction of the price actually paid by customers for a
product as discounts and allowances are given at various stages of the sales
process. Because few customers take advantage of all discounts, in analyzing a
product’s price waterfall, marketers must consider not only the amount of each
discount but also the percentage of sales to which it applies.
As customers vary in their use of discounts, net price can fall into a wide range relative
to list price.
Chapter 8 Promotion
289
Construction
To assess a product’s price waterfall, one must plot the price a customer will pay at each
stage of the waterfall, specifying potential discounts and allowances in the sequence in
which those are usually taken or applied. For example, broker commissions are generally applied after trade discounts.
Net Price: The actual average price paid for a product at a given stage in its distribution channel can be calculated as its list price, less discounts offered, with
each discount multiplied by the probability that it will be applied. When all discounts are considered, this calculation yields the product’s net price.
Net Price ($) = List Price ($) – [Discount A ($) * Proportion of Purchases on which Discount
A is Taken (%)] – [Discount B ($) * Proportion of Purchases on which
Discount B is Taken (%)] and so on . . .
Price Waterfall Effect (%) =
Net Price per Unit ($)
List Price per Unit ($)
EXAMPLE: Hakan manages his own firm. In selling his product, Hakan grants two discounts
or allowances. The first of these is a 12% discount on orders of more than 100 units. This is given on
50% of the firm’s business and appears on its invoicing system. Hakan also gives an allowance of 5%
for cooperative advertising. This is not shown on the invoicing system. It is completed in separate
procedures that involve customers submitting advertisements for approval. Upon investigation,
Hakan finds that 80% of customers take advantage of this advertising allowance.
The invoice price of the firm’s product can be calculated as the list price (50 Dinar per unit), less the
12% order size discount, multiplied by the chance of that discount being given (50%).
Invoice Price = List Price – [Discount * Proportion of Purchases on Which Discount Is Taken]
= 50 Dinar – [(50 * 12%) * 50%]
= 50 Dinar – 3 Dinar = 47 Dinar
The net price further reduces the invoice price by the average amount of the cooperative advertising
allowance granted, as follows:
Net Price = List Price – [Discount * Proportion of Purchases on Which Discount Is Taken]
– [Advertising Allowance * Proportion of Purchases on Which Ad Allowance Is
Taken] = 50 Dinar – [(50 * 12%) * 50%] – [(50 * 5%) * 80%] = 50 – 3 – 2
= 45 Dinar
To find the effect of the price waterfall, divide the net price by the list price.
Price Waterfall (%) =
290
MARKETING METRICS
45
50
= 90%
Data Sources, Complications, and Cautions
To analyze the impact of discounts, allowances, and the overall price waterfall effect,
marketers require full information about sales, in both revenue and unit volume terms,
at an individual product level, including not only those discounts and allowances that
are formally recorded in the billing system, but also those granted without appearing
on invoices.
The major challenge in establishing the price waterfall is securing product-specific data
at all of these various levels in the sales process. In all but the smallest businesses, this
is likely to be quite difficult, particularly because many discounts are granted on an
off-invoice basis, so they might not be recorded at a product level in a firm’s financial
system. Further complicating matters, not all discounts are based on list price. Cash
discounts, for example, are usually based on net invoice price.
Where discounts are known in theory, but the financial system doesn’t fully record
their details, the problem is determining how to calculate the price waterfall. Toward
that end, marketers need not only the amount of each discount, but also the percentage
of unit sales for which customers take advantage of that discount.
The typical business offers a number of discounts from list prices. Most of these serve
the function of encouraging particular customer behaviors. For example, trade discounts can encourage distributors and resellers to buy in full truckloads, pay invoices
promptly, and place orders during promotional periods or in a manner that smoothes
production. Over time, these discounts tend to multiply as manufacturers find it easier
to raise list price and add another discount than to eliminate discounts altogether.
Problems with discounts include the following:
■
■
■
Because it’s difficult to record discounts on a per-item basis, firms often record
them in aggregate. On this basis, marketers may see the total discounts provided but have difficulty allocating these to specific products. Some discounts are
offered on the total size of a purchase, exacerbating this problem. This increases
the challenge of assessing product profitability.
Once given, discounts tend to be sticky. It is hard to take them away from customers. Consequently, inertia often leaves special discounts in place, long after
the competitive pressures that prompted them are removed.
To the extent that discounts are not recorded on invoices, management often
loses track of them in decision-making.
As the Professional Pricing Society advises, when considering the price of a product,
“Look past the invoice price.”6
Chapter 8 Promotion
291
Related Metrics and Concepts
Deductions: Some “discounts” are actually deductions applied by a customer to an
invoice, adjusting for goods damaged in shipment, incorrect deliveries, late deliveries,
or in some cases, for products that did not sell as well as hoped. Deductions might not
be recorded in a way that can be analyzed, and they often are the subject of disputes.
Everyday Low Prices (EDLP): EDLP refers to a strategy of offering the same pricing level from period to period. For retailers, there is a distinction between buying at
EDLP and selling at EDLP. For example, some suppliers offer constant selling prices
to retailers but negotiate periods during which a product will be offered on deal with
display and other retail promotions. Rather than granting temporary price discounts
to retailers, suppliers often finance these programs through “market development
funds.”
HI-LO (High-Low): This pricing strategy constitutes the opposite of EDLP. In HI-LO
pricing, retailers and manufacturers offer a series of “deals” or “specials”—times during
which prices are temporary decreased. One purpose of HI-LO pricing and other temporary discounts is to realize price discrimination in the economic—not the legal—sense
of the term.
Price Discrimination and Tailoring
When firms face distinct and separable market segments with different willingness to
pay (price elasticities), charging a single price means that the firm will “leave money on
the table”—not capture the full consumer value.
There are three conditions for price tailoring to be profitable:
■
■
■
Segments must have different elasticities (willingness to pay), and/or marketers must have different costs of serving the segments (say shipping expenses)
and the incremental volume must be sufficiently large to compensate for the
reduction in margin.
Segments must be separable—that is, charging different prices does not just
result in transfer between segments (for example, your father cannot buy your
dinner and apply the senior citizen discount).
The incremental profit from price tailoring exceeds the costs of implementing multiple prices for the same product or service.
Price tailoring is clearly a euphemism for price discrimination. However, the latter term
is loaded with legal implications, and marketers understandably use it with caution.
292
MARKETING METRICS
When facing a total demand curve composed of identifiable segments with different
demand slopes, a marketer can use optimal pricing for each segment recognized, as
opposed to using the same price based upon aggregate demand. This is usually done by
■
■
■
Time: For example, subways or movie theaters charging a higher price during
rush or peak hour or products that are launched at a high price in the beginning, “skimming” profits from early adopters.
Geography: Such as international market divisions—different prices for different regions for DVDs, for example.
Tolerable discrimination: Identifying acceptable forms of segmentation, such
as discriminating between students or senior citizens and the general public.
Price differences cause gray markets; goods are imported from low-price to high-price
markets. Gray markets are common in some fashion goods and pharmaceuticals.
Caution: Regulations
Most countries have regulations that apply to price discrimination. As a marketer,
you should understand these regulations. In the U.S., the most important regulation is the Robinson-Patman Act. It is mainly intended to control price differences that might injure competition.7 We encourage you to visit the Federal Trade
Commission’s Web site (www.ftc.gov) for more information.
References and Suggested Further Reading
Abraham, M.M., and L.M. Lodish. (1990). “Getting the Most Out of Advertising and
Promotion,” Harvard Business Review, 68(3), 50.
Ailawadi, K., P. Farris, and E. Shames. (1999). “Trade Promotion: Essential to Selling Through
Resellers,” Sloan Management Review, 41(1), 83–92.
Christen, M., S. Gupta, J.C. Porter, R. Staelin, and D.R. Wittink. (1997). “Using Market-level
Data to Understand Promotion Effects in a Nonlinear Model,” Journal of Marketing Research
(JMR), 34(3), 322.
Roegner, E., M. Marn, and C. Zawada. (2005). “Pricing,” Marketing Management, Jan/Feb,
Vol. 14 (1).
Chapter 8 Promotion
293
This page intentionally left blank
9
ADVERTISING METRICS
Introduction
Key concepts covered in this chapter:
Advertising: Impressions, Gross Rating
Points, and Opportunities-to-See
Frequency Response Functions
Cost per Thousand Impressions
(CPM) Rates
Share of Voice
Reach/Net Reach and Frequency
Effective Reach and Effective Frequency
Advertising Elasticity of Demand
Advertising is the cornerstone of many marketing strategies. The positioning and communications conveyed by advertising often set the tone and timing for many other sales
and promotion efforts. Advertising is not only the defining element of the marketing
mix, but it is also expensive and notoriously difficult to evaluate. This is because it is
not easy to track the incremental sales associated with advertising decisions. For many
marketers, media metrics are particularly confusing. A command of the vocabulary
involved in this field is needed to work with media planners, buyers, and agencies. A
strong understanding of media metrics can help marketers ensure that advertising budgets are spent efficiently and directed toward a specific aim.
In this chapter, we discuss media metrics that reveal how many people may be exposed
to an advertising campaign, how often those people have an opportunity to see the ads,
and the cost of each potential impression. Toward that end, we introduce the vocabulary of advertising metrics, including such terms as impressions, exposures, OTS, rating
points, GRPs, net reach, effective frequency, and CPMs.
295
Metric
Construction
Considerations
Purpose
9.1
Impressions
An impression is
generated each time
an advertisement
is viewed. The
number of
impressions
achieved is a
function of an ad’s
reach (the number
of people seeing it),
multiplied by its
frequency (number
of times they see it).
As a metric,
impressions do not
account for quality
of viewings. In this
regard, a glimpse
will have less effect
than a detailed
study. Impressions
are also called
exposures and
opportunities-tosee (OTS).
To understand
how many times
an advertisement is
viewed.
9.1
Gross Rating
Points (GRPs)
Impressions
divided by the
number of
people in the
audience for an
advertisement.
Impressions
expressed in
relation to
population. GRPs
are cumulative
across media
vehicles, making
it possible to
achieve GRPs of
more than 100%.
Target Rating
Points (TRPs)
are measured in
relation to defined
target populations.
To measure
impressions in
relation to the
number of people
in the audience
for an advertising
campaign.
9.2
Cost per Thousand Cost of advertising
Impressions (CPM) divided by
impressions
generated (in
thousands).
CPM is a
measure of cost
per advertising
impression,
reckoning
impressions in
thousands. This
makes it easier
to work with the
resulting dollar
figures than would
be possible on the
basis of cost per
single impression.
To measure the
cost-effectiveness
of the generation of
impressions.
296
MARKETING METRICS
Metric
Construction
Considerations
Purpose
9.3
Net Reach
The number of
people who are
exposed to an
advertisement.
Equivalent to reach.
Measures unique viewers
of an advertisement.
Often best mapped on a
Venn diagram.
To measure the
breadth of an
advertisement’s
spread across a
population.
9.3
Average
Frequency
The average number
of times that an
individual is exposed
to an advertisement,
given that he or she
is indeed exposed to
the ad.
Frequency is measured
only among people who
have been exposed to
the advertisement under
study.
To measure
how often an
advertisement is
exposed to a given
population.
9.4
Frequency
Response
Functions
Linear: All
advertising
impressions are
equally impactful.
Linear model is often
To model the
unrealistic, especially for response of
complex products.
an audience
Threshold model is often to additional
used, as it is simple and exposures of an
advertisement.
intuitive.
Threshold: A
certain number of
impressions are
needed before an
advertising message
will sink in.
Learning curve: An
advertisement has
little impact at first
but gains force with
repetition and then
tails off as saturation
is achieved.
Learning curve models
often hypothesized,
but difficult to test
for accuracy. Simpler
models often work
as well.
Continues
Chapter 9 Advertising Metrics
297
Metric
Construction
Considerations
Purpose
9.5
Effective
Reach
Reach achieved
among individuals
who are exposed to
an advertisement
with a frequency
greater than or
equal to the effective
frequency.
The effective frequency
rate constitutes a
crucial assumption in
the calculation of this
metric.
To measure the
portion of an
audience that is
exposed to an
advertisement
enough times to be
influenced.
9.5
Effective
Frequency
The number of times
an individual must
see an advertisement
in order to respond at
the desired or target
level.
As a rule of thumb in
planning, marketers
often use an effective
frequency of 3. To the
extent that it promises
to have a significant
impact on campaign
results, this assumption
should be tested.
To determine
optimal exposure
levels for an
advertisement or
campaign, trading
the risk of overspending against
the risk of failing to
achieve the desired
impact.
9.6
Share of
Voice
Quantifies the
advertising
“presence” of a
brand, campaign, or
firm in relation to
total advertising in a
market.
Market definition is
central to meaningful
results. Impressions
or ratings represent a
conceptually strong
basis for share of
voice calculations.
Often, however, such
data are unavailable.
Consequently,
marketers use
spending, an input, as
a proxy for output.
To evaluate the
relative strength
of advertising
program within its
market.
298
MARKETING METRICS
9.1 Advertising: Impressions, Exposures,
Opportunities-To-See (OTS), Gross Rating Points
(GRPs), and Target Rating Points (TRPs)
Advertising impressions, exposures, and opportunities-to-see (OTS) all refer to
the same metric: an estimate of the audience for a media “insertion” (one ad) or
campaign.
Impressions = OTS = Exposures. In this chapter, we will use all these terms. It is
important to distinguish between “reach” (number of unique individuals exposed
to certain advertising) and “frequency” (the average number of times each such
individual is exposed).
Rating Point = Reach of a media vehicle as a percentage of a defined population (for
example, a television show with a rating of 2 reaches 2% of the population).
Gross Rating Points (GRPs) = Total Ratings achieved by multiple media vehicles
expressed in rating points (for example, advertisements on five television shows with
an average rating of 30% would achieve 150 GRPs).
Gross rating points are impressions expressed as a percentage of a defined population, and often total more than 100%. This metric refers to the defined population
reached rather than an absolute number of people. Although GRPs are used with a
broader audience, the term target rating points (TRPs) denotes a narrower definition of the target audience. For example, TRPs might consider a specific segment
such as youths aged 15 to 19, whereas GRPs might be based on the total TV viewing
population.
Purpose: To measure the audience for an advertisement.
Impressions, exposures, and opportunities-to-see (OTS) are the “atoms” of media
planning. Every advertisement released into the world has a fixed number of planned
exposures, depending on the number of individuals in its audience. For example, an
advertisement that appears on a billboard on the Champs-Élysées in central Paris will
have an estimated number of impressions, based on the flow of traffic from visitors and
locals. An advertisement is said to “reach” a certain number of people on a number
of occasions, or to provide a certain number of “impressions” or “opportunities-tosee.” These impressions or opportunities-to-see are thus a function of the number of
people reached and the number of times each such person has an opportunity to see
the advertisement.
Chapter 9 Advertising Metrics
299
Methodologies for estimating opportunities-to-see vary by type of media. In magazines,
for example, opportunities-to-see will not equal circulation because each copy of the
magazine may be read by more than one person. In broadcast media, it is assumed that
the quantified audience comprises those individuals available to hear or see an advertisement. In print and outdoor media, an opportunity-to-see might range from a brief
glance to a careful consideration. To illustrate this range, imagine you’re walking down
a busy street. How many billboard advertisements catch your eye? You may not realize
it, but you’re contributing to the impressions of several advertisements, regardless of
whether you ignore them or study them with great interest.
When a campaign involves several types of media, marketers may need to adjust their
measures of opportunities-to-see in order to maintain consistency and allow for comparability among the different media.
Gross rating points (GRPs) are related to impressions and opportunities-to-see. They
quantify impressions as a percentage of the population reached rather than in absolute
numbers of people reached. Target rating points (TRPs) express the same concept but
with regard to a more narrowly defined target audience.
Construction
Impressions, Opportunities-to-See (OTS), and Exposures: The number of
times a specific advertisement is available to be seen or otherwise exposed to
media audiences. This is an estimate of the audience for a media “insertion” (one
ad) or a campaign. Impressions = OTS = Exposures.
Impressions: The process of estimating reach and frequency begins with data that
sum all of the impressions from different advertisements to arrive at total “gross”
impressions.
Impressions (#) = Reach (#) * Average Frequency (#)
The same formula can be rearranged as follows to convey the average number of times
that an audience was given the opportunity to see an advertisement. Average frequency
is defined as the average number of impressions per individual “reached” by an advertisement or campaign.
Average Frequency (#) =
Impressions (#)
Reach (#)
Similarly, the reach of an advertisement—that is, the number of people with an opportunity to see the ad—can be calculated as follows:
Reach (#) =
300
MARKETING METRICS
Impressions (#)
Average Frequency (#)
Although reach can thus be quantified as the number of individuals exposed to an
advertisement or campaign, it can also be calculated as a percentage of the population.
In this text, we will distinguish between the two conceptualizations of this metric as
reach (#) and reach (%).
The reach of a specific media vehicle, which may deliver an advertisement, is often
expressed in rating points. Rating points are calculated as individuals reached by
that vehicle, divided by the total number of individuals in a defined population, and
expressed in “points” that represent the resulting percentage. Thus, a television program with a rating of 2 would reach 2% of the population.
The rating points of all the media vehicles that deliver an advertisement or campaign
can be summed, yielding a measure of the aggregate reach of the campaign, known as
gross rating points (GRPs).
Gross Rating Points (GRPs): The sum of all rating points delivered by the
media vehicles carrying an advertisement or campaign.
EXAMPLE: A campaign that delivers 150 GRPs might expose 30% of the population to an
advertisement at an average frequency of 5 impressions per individual (150 = 30 * 5). If 15 separate
“insertions” of the advertisement were used, a few individuals might be exposed as many as 15
times, and many more of the 30% reached would only have 1 or 2 opportunities-to-see (OTS).
Gross Rating Points (GRPs) (%) = Reach (%) * Average Frequency (#)
Gross Rating Points (GRPs) (%) =
Impressions (#)
Defined Population (#)
Target Rating Points (TRPs): The gross rating points delivered by a media
vehicle to a specific target audience.
EXAMPLE: A firm places 10 advertising insertions in a market with a population of 5 people.
The resulting impressions are outlined in the following table, in which “1” represents an opportunity-to-see, and “0” signifies that an individual did not have an opportunity to see a particular
insertion.
In this campaign, the total impressions across the entire population = 22.
Chapter 9 Advertising Metrics
301
Insertion
A
B
C
D
E
Impressions
Rating Points
(Impressions/
Population)
1
1
1
0
0
1
3
60
2
1
1
0
0
1
3
60
3
1
1
0
1
0
3
60
4
1
1
0
1
0
3
60
5
1
1
0
1
0
3
60
6
1
0
0
1
0
2
40
7
1
0
0
1
0
2
40
8
1
0
0
0
0
1
20
9
1
0
0
0
0
1
20
10
1
0
0
0
0
1
20
Totals
10
5
0
5
2
22
440
Individual
As insertion 1 generates impressions upon three of the five members of the population, it reaches
60% of that population, for 60 rating points. As insertion 6 generates impressions upon two of the
five members of the population, it reaches 40% of the population, for 40 rating points. Gross rating
points for the campaign can be calculated by adding the rating points of each insertion.
Gross Rating Points (GRPs) = Rating Points of Insertion 1 + Rating Points of Insertion 2 + etc.
= 440
Alternatively, gross rating points can be calculated by dividing total impressions by the size of the
population and expressing the result in percentage terms.
Gross Rating Points (GRPs) =
Impressions
Population
* 100% =
22
5
* 100% = 440
Target rating points (TRPs), by contrast, quantify the gross rating points achieved by an advertisement or campaign among targeted individuals within a larger population. For purposes of this
example, let’s assume that individuals A, B, and C comprise the targeted group. Individual A has
received 10 exposures to the campaign; individual B, 5 exposures; and individual C, 0 exposures.
Thus, the campaign has reached two out of three, or 66.67% of targeted individuals. Among those
302
MARKETING METRICS
reached, its average frequency has been 15/2, or 7.5. On this basis, we can calculate target rating
points by either of the following methods.
Target Rating Points (TRPs) = Reach (%) * Average Frequency
= 66.67% *
15
2
= 500
Target Rating Points (TRPs) =
Impressions (#)
Targets (#)
=
15
3
= 500
Data Sources, Complications, and Cautions
Data on the estimated audience size (reach) of a media vehicle are typically made available by media sellers. Standard methods also exist for combining data from different
media to estimate “net reach” and frequency. An explanation of these procedures is
beyond the scope of this book, but interested readers might want to consult a company
dedicated to tracking rating points, such as Nielsen (www.nielsen.com), for further
detail.
Two different media plans can yield comparable results in terms of costs and total
exposures but differ in reach and frequency measures. In other words, one plan can
expose a larger audience to an advertising message less often, while the other delivers
more exposures to each member of a smaller audience. For an example, please see
Table 9.1.
Table 9.1 Illustration of Reach and Frequency
Reach
Average Frequency* Total Exposures (Impressions, OTS)
Plan A
250,000
4
1,000,000
Plan B
333,333
3
1,000,000
*Average frequency is the average number of exposures made to each individual who has received at least
one exposure to a given advertisement or campaign. To compare impressions across media, or even within
classes of media, one must make a broad assumption: that there is some equivalency between the different types of impressions generated by each media classification. Nonetheless, marketers must still compare the “quality” of impressions delivered by different media.
Chapter 9 Advertising Metrics
303
Consider the following examples: A billboard along a busy freeway and a subway
advertisement can both yield the same number of impressions. Whereas the subway
advertisement has a captive audience, members of the billboard audience are generally driving and concentrating on the road. As this example demonstrates, there may
be differences in the quality of impressions. To account for these differences, media
optimizers apply weightings to different media vehicles. When direct response data
are available, they can be used to evaluate the relative effectiveness and efficiency of
impression purchases in different media. Otherwise, this weighting might be a matter
of judgment. A manager might believe, for example, that an impression generated by
a TV commercial is twice as effective as one made by a magazine print advertisement.
Similarly, marketers often find it useful to define audience sub-groups and generate
separate reach and frequency statistics for each. Marketers might weight sub-groups
differently in the same way that they weight impressions delivered through different
media.1 This helps in evaluating whether an advertisement reaches its defined customer
groups.
When calculating impressions, marketers often encounter an overlap of people who see
an advertisement in more than one medium. Later in this text, we will discuss how to
account for such overlap and estimate the percentage of people who are exposed to an
advertisement multiple times.
9.2 Cost per Thousand Impressions (CPM) Rates
Cost per thousand impressions (CPM) is the cost per thousand advertising impressions. This metric is calculated by dividing the cost of an advertising placement by
the number of impressions (expressed in thousands) that it generates.
Cost per Thousand Impressions (CPM) ($) =
Advertising Cost ($)
Impressions Generated (# in Thousands)
CPM is useful in comparing the relative efficiency of different advertising opportunities or media and in evaluating the costs of overall campaigns.
Purpose: To compare the costs of advertising campaigns within and across
different media.
A typical advertising campaign might try to reach potential consumers in multiple locations and through various media. The cost per thousand impressions (CPM) metric
enables marketers to make cost comparisons between these media, both at the planning
stage and during reviews of past campaigns.
304
MARKETING METRICS
Marketers calculate CPM by dividing advertising campaign costs by the number of
impressions (or opportunities-to-see) that are delivered by each part of the campaign.
As the impression counts are generally sizable, marketers customarily work with the
CPM impressions. Dividing by 1,000 is an industry standard.
Cost per Thousand Impressions (CPM): The cost of a media campaign, relative to its success in generating impressions or opportunities-to-see.
Construction
To calculate CPM, marketers first state the results of a media campaign (gross impressions) in thousands. Second, they divide that result into the relevant media cost:
Cost per Thousand Impressions (CPM) ($) =
Advertising Cost ($)
Impressions Generated (# in Thousands)
EXAMPLE: An advertising campaign costs $4,000 and generates 120,000 impressions. On this
basis, CPM can be calculated as follows:
Cost per Thousand Impressions =
=
=
Advertising Cost
Impressions Generated (thousands)
$4,000
(120,000/1,000)
$4,000
120
= $33.33
Data Sources, Complications, and Cautions
In an advertising campaign, the full cost of the media purchased can include agency
fees and production of creative materials, in addition to the cost of media space or
time. Marketers also must have an estimate of the number of impressions expected
or delivered in the campaign at an appropriate level of detail. Internet marketers (see
Chapter 10, “Online, Email, and Mobile Metrics”) often can easily access these data.
CPM is only a starting point for analysis. Not all impressions are equally valuable.
Consequently, it can make good business sense to pay more for impressions from some
sources than from others.
In calculating CPM, marketers should also be concerned with their ability to capture
the full cost of advertising activity. Cost items typically include the amount paid to a
Chapter 9 Advertising Metrics
305
creative agency to develop advertising materials, amounts paid to an organization that
sells media, and internal salaries and expenses related to overseeing the advertisement.
Related Metrics and Concepts
Cost per Point (CPP): The cost of an advertising campaign, relative to the rating points
delivered. In a manner similar to CPM, cost per point measures the cost per rating
point for an advertising campaign by dividing the cost of the advertising by the rating
points delivered.
9.3 Reach, Net Reach, and Frequency
Reach is the same as net reach; both of these metrics quantify the number or percentage of individuals in a defined population who are exposed to at least one exposure to an advertisement. Frequency measures the average number of times that each
such individual sees the advertisement.
Impressions (#) = Reach (#) * Frequency (#)
Net reach and frequency are important concepts in describing an advertising campaign. A campaign with a high net reach and low frequency runs the danger of
being lost in a noisy advertising environment. A campaign with low net reach but
high frequency can over-expose some audiences and miss others entirely. Reach and
frequency metrics help managers adjust their advertising media plans to fit their
marketing strategies.
Purpose: To separate total impressions into the number of people reached
and the average frequency with which those individuals are exposed
to advertising.
To clarify the difference between reach and frequency, let’s review what we learned
in Section 9.1. When impressions from multiple insertions are combined, the results
are often called “gross impressions” or “total exposures.” When total impressions are
expressed as a percentage of the population, this measure is referred to as gross rating
points (GRPs). For example, suppose a media vehicle reaches 12% of the population.
That vehicle will have a single-insertion reach of 12 rating points. If a firm advertised in
10 such vehicles, it would achieve 120 GRPs.
306
MARKETING METRICS
Now, let’s look at the composition of these 120 GRPs. Suppose we know that the 10
advertisements had a combined net reach of 40% and an average frequency of 3. Then
their gross rating points might be calculated as 40 * 3 = 120 GRPs.
EXAMPLE: A commercial is shown once in each of three time slots. Nielsen keeps track of
which households have an opportunity to see the advertisement. The commercial airs in a market
with only five households: A, B, C, D, and E. Time slots 1 and 2 both have a rating of 60 because 60%
of the households view them. Time slot 3 has a rating of 20.
Time Slot
Households with
Opportunity-to-See
Households with no
Opportunity-to-See
Rating Points of
Time Slot
1
ABE
CD
60
2
ABC
DE
60
3
A
BCDE
20
GRP
GRP =
Impressions
Population
=
7
5
140
= 140 (%)
The commercial is seen by households A, B, C, and E, but not D. Thus, it generates impressions in
four out of five households, for a reach (%) of 80%. In the four households reached, the commercial
is seen a total of seven times. Thus, its average frequency can be calculated as 7/4, or 1.75. On this
basis, we can calculate the campaign’s gross rating points as follows:
GRP = Reach (%) * Average Frequency (#) =
4
5
*
7
4
= 80% * 1.75 = 140 (%)
Unless otherwise specified, simple measures of overall audience size (such as GRPs
or impressions) do not differentiate between campaigns that expose larger audiences
fewer times and those that expose smaller audiences more often. In other words, these
metrics do not distinguish between reach and frequency.
Reach, whether described as “net reach” or simply “reach,” refers to the unduplicated
audience of individuals who have been exposed at least once to the advertising in question. Reach can be expressed as either the number of individuals or the percentage of
the population that has seen the advertisement.
Reach: The number of people or percent of population exposed to an
advertisement.
Chapter 9 Advertising Metrics
307
Frequency is calculated by dividing gross impressions by reach. Frequency is equal to
the average number of exposures received by individuals who have been exposed to at
least one impression of the advertising in question. Frequency is calculated only among
individuals who have been exposed to this advertising. On this basis: Total Impressions
= Reach * Average Frequency.
Average Frequency: The average number of impressions per reached individual.
Media plans can differ in reach and frequency but still generate the same number of
total impressions.
Net Reach: This term is used to emphasize the fact that the reach of multiple advertising placements is not calculated through the gross addition of all
individuals reached by each of those placements. Occasionally, the word “net”
is eliminated, and the metric is called simply reach.
EXAMPLE: Returning to our prior example of a 10-insertion media plan in a market with a
population of five people, we can calculate the reach and frequency of the plan by analyzing the following data. As previously noted, in the following table, “1” represents an opportunity-to-see, and
“0” signifies that an individual did not have an opportunity to see a particular insertion.
Insertion
A
B
C
D
E
Impressions
Rating Points
(Impressions/
Population)
1
1
1
0
0
1
3
60
2
1
1
0
0
1
3
60
3
1
1
0
1
0
3
60
4
1
1
0
1
0
3
60
5
1
1
0
1
0
3
60
6
1
0
0
1
0
2
40
7
1
0
0
1
0
2
40
8
1
0
0
0
0
1
20
9
1
0
0
0
0
1
20
10
1
0
0
0
0
1
20
Totals
10
5
0
5
2
22
440
Individual
308
MARKETING METRICS
Reach is equal to the number of people who saw at least one advertisement. Four of the five people
in the population (A, B, D, and E) saw at least one advertisement. Consequently, reach (#) = 4.
Average Frequency =
Impressions
Reach
=
22
4
= 5.5
When multiple vehicles are involved in an advertising campaign, marketers need information about the overlap among these vehicles as well as sophisticated mathematical
procedures in order to estimate reach and frequency. To illustrate this concept, the
following two-vehicle example can be useful. Overlap can be represented by a graphic
known as a Venn diagram (see Figure 9.1).
Population needs to be excluded from the
reach of the second exposure to prevent
double counting.
Population with No Exposure
1 Exposure
Advertisement A
Figure 9.1
1 Exposure
2
Advertisement B
Venn Diagram Illustration of Net Reach
EXAMPLE: As an illustration of overlap effects, let’s look at two examples. Aircraft International
magazine offers 850,000 impressions for one advertisement. A second magazine, Commercial
Flying Monthly, offers 1 million impressions for one advertisement.
Chapter 9 Advertising Metrics
309
Example 1: Marketers who place advertisements in both magazines should not expect to reach 1.85
million readers. Suppose that 10% of Aircraft International readers also read Commercial Flying
Monthly. On this basis, net reach = (850,000 * .9) + 1,000,000 = 1,765,000 unique individuals. Of
these, 85,000 (10% of Aircraft International readers) have received two exposures. The remaining
90% of Aircraft International readers have received only one exposure. The overlap between two
different media types is referred to as external overlap.
Example 2: Marketers often use multiple insertions in the same media vehicle (such as the July and
August issues of the same magazine) to achieve frequency. Even if the estimated audience size is
the same for both months, not all of the same people will read the magazine each month. For purposes of this example, let’s assume that marketers place insertions in two different issues of Aircraft
International, and that only 70% of readers of the July issue also read the August issue. On this basis,
net reach is not merely 850,000 (the circulation of each issue of Aircraft International) because the
groups viewing the two insertions are not precisely the same. Likewise, net reach is not 2 * 850,000,
or 1.7 million, because the groups viewing the two insertions are also not completely disparate.
Rather, net reach = 850,000 + (850,000 * 30%) = 1,105,000.
The reason: Thirty percent of readers of the August issue did not read the July issue and so did
not have the opportunity to see the July insertion of the advertisement. These readers—and only
these readers—represent incremental viewers of the advertisement in August, and so they must
be added to net reach. The remaining 70% of August readers were exposed to the advertisement
twice. Their total represents internal overlap or duplication.
Data Sources, Complications, and Cautions
Although we’ve emphasized the importance of reach and frequency, the impressions
metric is typically the easiest of these numbers to establish. Impressions can be aggregated on the basis of data originating from the media vehicles involved in a campaign.
To determine net reach and frequency, marketers must know or estimate the overlap
between audiences for different media, or for the same medium at different times. It is
beyond the capability of most marketers to make accurate estimates of reach and frequency without access to proprietary databases and algorithms. Full-service advertising
agencies and media buying companies typically offer these services.
Assessing overlap is a major challenge. Although overlap can be estimated by performing customer surveys, it is difficult to do this with precision. Estimates based on managers’ judgment occasionally must suffice.
310
MARKETING METRICS
9.4 Frequency Response Functions
Frequency response functions help marketers to model the effectiveness of multiple
exposures to advertising. We discuss three typical assumptions about how people
respond to advertisements: linear response, learning curve response, and threshold
response.
In a linear response model, people are assumed to react equally to every exposure
to an advertisement. The learning curve response model assumes that people are
initially slow to respond to an advertisement and then respond more quickly for a
time, until ultimately they reach a point at which their response to the message tails
off. In a threshold response function, people are assumed to show little response
until a critical frequency level is reached. At that point, their response immediately
rises to maximum capacity.
Frequency response functions are not technically considered metrics. Understanding
how people respond to the frequency of their exposure to advertising, however, is
a vital part of media planning. Response models directly determine calculations of
effective frequency and effective reach, metrics discussed in Section 9.5.
Purpose: To establish assumptions about the effects
of advertising frequency.
Let’s assume that a company has developed a message for an advertising campaign, and
that its managers feel confident that appropriate media for the campaign have been
selected. Now they must decide: How many times should the advertisement be placed?
The company wants to buy enough advertising space to ensure that its message is effectively conveyed, but it also wants to ensure that it doesn’t waste money on unnecessary
impressions.
To make this decision, a marketer will have to make an assumption about the
value of frequency. This is a major consideration: What is the assumed value of repetition in advertising? Frequency response functions help us to think through the value
of frequency.
Frequency Response Function: The expected relationship between advertising
outcomes (usually in unit sales or dollar revenues) and advertising frequency.
There are a number of possible models for the frequency response functions used in
media plans. A selection among these for a particular campaign will depend on the
product advertised, the media used, and the judgment of the marketer. Three of the
most common models are described next.
Chapter 9 Advertising Metrics
311
Linear Response: The assumption behind a linear response function is that each
advertising exposure is equally valuable, regardless of how many other exposures
to the same advertising have preceded it.
Learning Curve Response: The learning or S curve model rests on the assumption that a consumer’s response to advertising follows a progression: The first
few times an advertisement is shown, it does not register with its intended audience. As repetition occurs, the message permeates its audience and becomes more
effective as people absorb it. Ultimately, however, this effectiveness declines, and
diminishing returns set in. At this stage, marketers believe that individuals who
want the information already have it and can’t be influenced further; others simply are not interested.
Threshold Response: The assumption behind this model is that advertising
has no effect until its exposure reaches a certain level. At that point, its message
becomes fully effective. Beyond that point, further advertising is unnecessary and
would be wasted.
These are three common ways to value advertising frequency. Any function that accurately describes the effect of a campaign can be used. Typically, however, only one function will apply to a given situation.
Construction
Frequency response functions are most useful if they can be used to quantify the effects
of incremental frequency. To illustrate the construction of the three functions described
in this section, we have tabulated several examples.
Tables 9.2 and 9.3 show the assumed incremental effects of each exposure to a certain
advertising campaign. Suppose that the advertisement will achieve maximum effect
(100%) at eight exposures. By analyzing this effect in the context of various response
functions, we can determine when and how quickly it takes hold.
Under a linear response model, each exposure below the saturation point generates
one-eighth, or 12.5%, of the overall effect.
The learning curve model is more complex. In this function, the incremental effectiveness of each exposure increases until the fourth exposure and declines thereafter.
Under the threshold response model, there is no effect until the fourth exposure. At
that point, however, 100% of the benefit of advertising is immediately realized. Beyond
that point, there is no further value to be obtained through incremental advertising.
Subsequent exposures are wasted.
The effects of these advertising exposures are tabulated cumulatively in Table 9.3. In
this display, maximum attainable effectiveness is achieved when the response to advertising reaches 100%.
312
MARKETING METRICS
Table 9.2 Example of the Effectiveness of Advertising
Exposure Frequency
Linear
Learning or S Curve
Threshold Value
1
0.125
0.05
0
2
0.125
0.1
0
3
0.125
0.2
0
4
0.125
0.25
1
5
0.125
0.2
0
6
0.125
0.1
0
7
0.125
0.05
0
8
0.125
0.05
0
Table 9.3 Assumptions: Cumulative Advertising Effectiveness
Exposure Frequency
Linear
Learning or S Curve
Threshold Value
1
12.5%
5%
0%
2
25.0%
15%
0%
3
37.5%
35%
0%
4
50.0%
60%
100%
5
62.5%
80%
100%
6
75.0%
90%
100%
7
87.5%
95%
100%
8
100.0%
100%
100%
We can plot cumulative effectiveness against frequency under each model (see
Figure 9.2). The linear function is represented by a simple straight line. The Threshold
assumption rises steeply at four exposures to reach 100%. The cumulative effects of the
learning curve model trace an S-shaped curve.
Frequency Response Function; Linear: Under this function, the cumulative
effect of advertising (up to the saturation point) can be viewed as a product of the
frequency of exposures and effectiveness per exposure.
Chapter 9 Advertising Metrics
313
Frequency Response Function; Linear (I) = Frequency (#) * Effectiveness per Exposure (I)
Frequency Response Function; Learning Curve: The learning curve function
can be charted as a non-linear curve. Its form depends on the circumstances of a
particular campaign, including selection of advertising media, target audience,
and frequency of exposures.
Frequency Response Function; Threshold: The threshold function can be
expressed as a Boolean “if” statement, as follows:
Frequency Response Function; Threshold Value (I) = If (Frequency (#) ≥ Threshold (#), 1, 0)
Stated another way: In a threshold response function, if frequency is greater than or
equal to the threshold level of effectiveness, then the advertising campaign is 100%
effective. If frequency is less than the threshold, there is no effect.
Conceptions of Advertising Effectiveness
100.0%
90.0%
Cumulative Effectiveness
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1
2
3
4
5
6
7
Frequency of Exposure
Linear
Figure 9.2
314
Learning or S-Curve
Threshold Value
Illustration of Cumulative Advertising Effectiveness
MARKETING METRICS
8
Data Sources, Complications, and Cautions
A frequency response function can be viewed as the structure of assumptions made
by marketers in planning for the effects of an advertising campaign. In making these
assumptions, a marketer’s most useful information can be derived from an analysis of
the effects of prior ad campaigns. Functions validated with past data, however, are most
likely to be accurate if the relevant circumstances (such as media, creative, price, and
product) have not significantly changed.
In comparing the three models discussed in this section, the linear response function has
the benefit of resting on a simple assumption. It can be unrealistic, however, because it is
hard to imagine that every advertising exposure in a campaign will have the same effect.
The learning curve has intuitive appeal. It seems to capture the complexity of life better than a linear model. Under this model, however, challenges arise in defining and
predicting an advertisement’s effectiveness. Three questions emerge: At what point
does the curve begin to ramp up? How steep is the function? When does it tail off? With
considerable research, marketers can make these estimates. Without it, however, there
will always be the concern that the learning curve function provides a spurious level of
accuracy.
Any implementation of the threshold response function will hinge on a firm’s estimate
of where the threshold lies. This will have important ramifications. If the firm makes a
conservative estimate, setting the tipping point at a high number of exposures, it may
pay for ineffective and unneeded advertising. If it sets the tipping point too low, however, it may not buy enough advertising media, and its campaign may fail to achieve the
desired effect. In implementation, marketers may find that there is little practical difference between using the threshold model and the more complicated learning curve.
Related Metrics and Concepts
Wear-in: The frequency required before a given advertisement or campaign achieves a
minimum level of effectiveness.
Wear-out: The frequency at which a given advertisement or campaign begins to lose
effectiveness or even yield a negative effect.
Chapter 9 Advertising Metrics
315
9.5
Effective Reach and Effective Frequency
The concept of effective frequency rests on the assumption that for an advertisement
or campaign to achieve an appreciable effect, it must attain a certain number of
exposures to an individual within a specified time period.
Effective reach is defined as the number of people or the percentage of the audience
that receives an advertising message with a frequency equal to or greater than the
effective frequency. That is, effective reach is the population receiving the “minimum” effective exposure to an advertisement or campaign.
Purpose: To assess the extent to which advertising audiences are being
reached with sufficient frequency.
Many marketers believe their messages require repetition to “sink in.” Advertisers,
like parents and politicians, therefore repeat themselves. But this repetition must be
monitored for effectiveness. Toward that end, marketers apply the concepts of effective
frequency and effective reach. The assumptions behind these concepts run as follows:
The first few times people are exposed to an ad, it may have little effect. It is only when
more exposures are achieved that the message begins to influence its audience.
With this in mind, in planning and executing a campaign, an advertiser must determine
the number of times that a message must be repeated in order to be useful. This number
is the effective frequency. In concept, this is identical to the threshold frequency in the
threshold response function discussed in Section 9.4. A campaign’s effective frequency
will depend on many factors, including market circumstances, media used, type of ad,
and campaign. As a rule of thumb, however, an estimate of three exposures per purchase cycle is used surprisingly often.
Effective Frequency: The number of times a certain advertisement must
be exposed to a particular individual in a given period to produce a desired
response.
Effective Reach: The number of people or the percentage of the audience that
receives an advertising message with a frequency equal to or greater than the
effective frequency.
Construction
Effective reach can be expressed as the number of people who have seen a particular
advertisement or the percentage of the population that has been exposed to that advertisement at a frequency greater than or equal to the effective frequency.
316
MARKETING METRICS
Effective Reach (#, %) = Individuals Reached with Frequency Equal to or
Greater Than Effective Frequency
EXAMPLE: An advertisement on the Internet was believed to need three exposures before its
message would sink in. Population data showed the distribution in Table 9.4.
Table 9.4
Number of Exposures of Advertisement
Number of Exposures
Population
0
140,000
1
102,000
2
64,000
3
23,000
4 or more
11,000
Total
340,000
Because the effective frequency is 3, only those who have seen the advertisement three or more
times have been effectively reached. The effective reach is thus 23,000 + 11,000 = 34,000.
In percentage terms, the effective reach of this advertisement is 34,000/340,000 = 10% of the
population.
Data Sources, Complications, and Cautions
The Internet has provided a significant boost to data gathering in this area. Although
even online campaigns can’t be totally accurate with regard to the number of advertisements served to each customer, data on this question from online campaigns are far
superior to those available in most other media.
Where data can’t be tracked electronically, it’s difficult to know how many times a
customer has been in a position to see an advertisement. Under these circumstances,
marketers make estimates on the basis of known audience habits and publicly available
resources, such as TV ratings.
Although test markets and split-cable experiments can shed light on the effects of
advertising frequency, marketers often lack comprehensive, reliable data on this question. In these cases, they must make—and defend—assumptions about the frequency
Chapter 9 Advertising Metrics
317
needed for an effective campaign. Even where good historical data are available, media
planning should not rely solely on past results because every campaign is different.
Marketers must also bear in mind that effective frequency attempts to quantify the
average customer’s response to advertising. In practice, some customers will need more
information and exposure than others.
9.6 Share of Voice
Share of voice quantifies the advertising “presence” that a specific product or brand
enjoys. It is calculated by dividing the brand’s advertising by total market advertising, and it is expressed as a percentage.
Share of Voice (%) =
Brand Advertising ($, #)
Total Market Advertising ($, #)
For purposes of share of voice, there are at least two ways to measure “advertising”:
in terms of dollar spending; or in unit terms, through impressions or gross rating
points (GRPs). By any of these measures, share of voice represents an estimate of a
company’s advertising, as compared to that of its competitors.
Purpose: To evaluate the comparative level of advertising committed
to a specific product or brand.
Advertisers want to know whether their messages are breaking through the “noise” in
the commercial environment. Toward that end, share of voice offers one indication of
a brand’s advertising strength, relative to the overall market.
There are at least two ways to calculate share of voice. The classic approach is to divide
a brand’s advertising dollar spend by the total advertising spend in the marketplace.
Alternatively, share of voice can be based on the brand’s share of GRPs, impressions,
effective reach, or similar measures (see earlier sections in this chapter for more details
on basic advertising metrics).
Construction
Share of Voice: The percentage of advertising in a given market that a specific
product or brand enjoys.
Share of Voice (%) =
318
MARKETING METRICS
Brand Advertising ($, #)
Total Market Advertising ($, #)
Data Sources, Complications, and Cautions
When calculating share of voice, a marketer’s central decision revolves around defining the boundaries of the market. One must ensure that these are meaningful to the
intended customer. If a firm’s objective is to influence Internet users, for example, it
would not be appropriate to define advertising presence solely in terms of print media.
Share of voice can be computed at a company level, but brand- and product-level
calculations are also common.
In executing this calculation, a company should be able to measure its total advertising spend fairly easily. Determining the ad spending for the market as a whole can be
fraught with difficulty, however. Complete accuracy will probably not be attainable.
It is important, however, that marketers take account of the major players in their
market. External sources such as annual reports and press clippings can shed light on
competitors’ ad spending. Publications such as Leading National Advertisers (LNA)
can also provide useful data. These services sell estimates of competitive purchases of
media space and time. They generally do not report actual payments for media, however. Instead, costs are estimated on the basis of the time and space purchased and on
published “rate cards” that list advertised prices. In using these estimates, marketers
must bear in mind that rate cards rarely cite the discounts available in buying media.
Without accounting for these discounts, published media spending estimates can be
inflated. Marketers are advised to deflate them by the discount rates they themselves
receive on advertising.
A final caution: Some marketers might assume that the price of advertising is equal to
the value of that advertising. This is not necessarily the case. With this in mind, it can
be useful to augment a dollar-based calculation of share of voice with one based on
impressions.
9.7 Advertising Elasticity of Demand
In this section we examine advertising elasticity, which represents the responsiveness
of sales to advertising.
Advertising Elasticity of Demand, AED (I)
=
Change in Quantity Demanded (%)
Change in Spending on Advertising (%)
We also discuss how the optimal level of advertising spending, as given by the
Dorfman-Steiner Theorem, relates to contribution margins, and the advertising/sales
ratio.
Chapter 9 Advertising Metrics
319
Purpose: Understanding the Responsiveness of Demand to Advertising.
In Chapter 7, “Pricing Strategy,” we considered the price elasticity of demand. This represents how responsive consumer demand, i.e., sales, is to a change in price. Advertising
elasticity of demand is a similar concept that represents the change in consumer demand
from an increase or decrease in advertising. Price elasticities are almost always negative
and less than -1.0, meaning a given percentage change in price will result in a greater
percentage change in sales, in the opposite direction of the price change. Advertising
elasticities are almost always positive and less than 1.0; a given percentage increase in
advertising spending will cause an increase in sales, but the percentage increase in sales
will be less than the percentage increase in advertising.
Armed with an estimate of consumers’ responsiveness to advertising, we can calculate
the profit-maximizing amount to spend on advertising. We can do this by comparing
the incremental contribution margin resulting from the change in sales revenue caused
by the change in advertising spending and compare this incremental contribution margin generated by the advertising to the marginal cost of the advertising.
Construction
The advertising elasticity of demand is simply the change in demand from a change in
advertising spending. This can be calculated by fitting an equation estimating advertising response as a function of advertising spending. To do this you will probably use
historical data, the results of an advertising test, or both.
The formula for advertising elasticity of demand is:
Advertising Elasticity of Demand—AED (I)
=
Change in Quantity Demanded (%)
Change in Spending on Advertising (%)
Unlike price elasticity of demand, advertising elasticity will almost certainly be positive
as increased spending on advertising should lead to greater demand. Greater elasticity
means that demand is more responsive to advertising. When elasticity is higher, spending more on advertising will be relatively more advantageous than when elasticity is
lower.
Typically, managers fit a constant elasticity model to data. The fitted equation would
contain an estimate of advertising elasticity (the slope coefficient in the regression of
ln(Sales) on ln(Advertising)). However, when only two points are available, the same
320
MARKETING METRICS
formula we noted earlier in the price elasticity section of Chapter 7 could be applied to
estimate the point elasticity of demand at current spending levels. This point elasticity
can be helpful in determining whether the firm is over- or under-spending, as we shall
show.
Where D1 is initial demand (sales in dollars) and D2 is demand after change in advertising. A1 is the initial level of advertising spending and A2 the level of advertising
spending after the change.
EXAMPLE: MCS Associates sells refilled printer ink cartridges. They do all of their advertising online, mainly through banner ads. MCS’s marketer wants to know the advertising elasticity of
demand and runs a test using different levels of advertising spend in two periods that are thought to
be comparable. The marketer is thus willing to assume that any difference in sales is caused by the
difference in advertising spending.
Let us assume that advertising elasticity of demand is constant.
Period 1: Advertising Spending (A1) was $10,000 and Quantity Demanded (D1) was $300,000
Period 2: Advertising Spending (A2) was $10,500 and Quantity Demanded (D2) was $303,000
Advertising elasticity of demand can be assessed by the impact of the increased advertising between
periods A and B. Given constant elasticity we must use the following formula.
= 0.204
What Should You Spend On Advertising?
After we know what the advertising elasticity of demand is, the key question becomes:
“are we spending too much, too little, or about the right amount?” An academic paper
from 1954 gives an answer to this question and introduced what is known as the
Chapter 9 Advertising Metrics
321
Dorfman-Steiner Theorem. This theorem gives an optimal level of advertising compared to sales given the contribution margin of the firm.
The contribution margin of the firm matters because some firms and industries have
relatively high contribution margins. In such a case most of the value generated by
increased demand (i.e., sales) is captured by the firm and so the firm is willing to
spend more on advertising to stimulate demand. When a firm has low contribution
margins, the firm requires a greater boost in sales to justify any increased spending on
advertising.
Dorfman-Steiner Theorem: The optimal level of advertising spending comparing sales
is as follows.
Advertising ($)
Sales ($)
=
Price ($) – Cost to Produce ($)
Price ($)
* AED
or Contribution Margin ($) * AED (I)
EXAMPLE: MCS Associates sells each cartridge at $10 and it costs $7.50 in variable costs to
acquire the used cartridges, fill them, accept and dispatch the order. Thus the contribution margin
is 25%. Assume that the advertising elasticity of demand (AED) is as above, a constant elasticity of
0.204.
The optimal level of advertising with respect to sales will be when Advertising ($)/Sales ($) equals
25%*AED, so 0.204*0.25 = 0.051. This implies 5.1% of sales should be spent on advertising by MCS
Associates.
When MCS was spending $10k on advertising and had sales of $300k, this is an advertising to sales
ratio of 3.3%. This gave a profit of $65,000 because Sales ($300,000) * Contribution Margin (25%) =
$75,000 less Advertising Spending ($10,000).
The Dorman-Steiner Theorem suggests the firm was under-advertising and would benefit from
advertising more heavily. The extra contribution from sales would outweigh the additional advertising costs. In this example, the optimal level of advertising is at an advertising to sales ratio of
5.1%, which is around $17,055 spent on advertising. This generates $334,507 in sales and in such a
scenario profit equals $66,572.
Table 9.5 illustrates an interesting effect that is called the flat-maximum principle. This states
that profit is often similar over a wide range of advertising spending. Increasing advertising from
$10,000 to $17,055 is a 71% increase in advertising spending but profits only increase by 2%. The
extra contribution generated by the advertising is not much more than the extra cost of the advertising. The good news for managers is that this means if your advertising spending isn’t massively
far off optimal you won’t lose too much from not being at the perfect level. Of course, a less opti-
322
MARKETING METRICS
mistic interpretation is that constant elasticities may be only rough estimates of sales response to
advertising.
Table 9.5: Advertising and Profits
Advertising
Spending ($)
Sales ($)
Contribution
Margin ($)
Profit ($)
Advertising to
Sales Ratio
=Scaling Factor =Sales *
* Advertising
Contribution
Spend ^AED
Margin(%)
=Contribution Margin
Minus Advertising
Advertising /
Spending
Sales
$10,000
$300,000
$75,000
$65,000
3.33%
$13,000
$316,489
$79,122
$66,122
4.11%
$16,000
$330,179
$82,545
$66,545
4.85%
$17,055
$334,507
$83,627
$66,572
5.10%
$19,000
$341,956
$85,489
$66,489
5.56%
$22,000
$352,335
$88,084
$66,084
6.24%
What If Advertising Is An Investment?
Advertising is typically treated as an expense. The logic of this suggests that the impact
of advertising on sales should be felt entirely at the time that the money is spent, i.e.,
it drives immediate sales. This misses the purpose of much advertising, such as that
designed to build a brand, which will drive more sales in future periods. Often advertising has a carryover effect—yesterday’s spending still has benefits for us today and will
have benefits for us tomorrow. How is the optimal level of advertising impacted by the
fact that advertising effects can carry on from one period to the next?
The interesting result is that the optimal advertising to spending ratio is unchanged
by carryover effects given the assumption of constant elasticity. When advertising has
greater impact in the long term because of carryover effects, the level of total sales rises.
This means that the optimal level of advertising also rises, but the ideal ratio between
the two remains the same. The key thing to remember is that the amount you should
spend on advertising rises where advertising has a greater long term effect but the optimal level of advertising still has the same ratio to the level of sales. However, our main
point is that advertising elasticities, contribution margins, and advertising-to-sales
ratios are linked in a way that can give managers insights into whether the firms are
spending too much or too little.
Chapter 9 Advertising Metrics
323
Data Sources, Complications and Cautions
A manager will need to be able to map out (or at least predict) demand at various levels
of advertising to use advertising elasticity of demand. This can be challenging for many
managers who may not have the right data or be able to run tests to estimate elasticity.
Applying the Dorfman-Steiner Theorem also assumes that the variables behave in a
predictable fashion. To use advertising elasticity to assess the appropriate advertising
to sales ratio it is often necessary to assume that elasticity is constant. This is the sort
of assumption that is often easier for academic economists to make than for practicing
managers. As such, the “scientific” determination of advertising to sales ratios has been
less widely used than the apparent power of the technique might suggest.
If the effect of advertising is not smooth the models are much harder to apply. For
instance, the impact of advertising spending might be lumpy and you need to reach
spending thresholds before advertising has any impact. In such cases raising advertising spending a little may not increase sales at all. Raising spending by a large amount,
however, might, for example allow a new media channel to be used, and so substantially
increase sales.
The Dorfman-Steiner result best applies to stable markets with well established brands.
When categories are forming (or undergoing fundamental change), however, brands
often spend more than is immediately profitable (even after considering carryover
effects) in battling for position (standards wars, for example).
Advertising elasticity and optimal advertising can assist thinking about advertising
effectiveness. The idea is to help managers approach the problem of how much to
advertise rather than provide them with “the answer,” which must be applied slavishly.
References and Suggested Further Reading
Dorfman, Robert and Peter O. Steiner. (1954) Optimal Advertising and Optimal Quality.
American Economic Review 44, 826-36.
Farris, Paul W., David Reibstein, and Ervin Shames. (1998). “Advertising Budgeting: A Report
from the Field,” monograph, New York: American Association of Advertising Agencies.
Forrester, J.W. (1959). “ADVERTISING: A Problem in Industrial Dynamics,” Harvard Business
Review, 37(2), 100.
Tellis, G.J., and D.L. Weiss. (1995). “Does TV Advertising Really Affect Sales? The Role of
Measures, Models, and Data Aggregation,” Journal of Advertising, 24(3), 1.
Tull, Donald S., Van R. Wood, Dale Duhan, Tom Gillpatrick, Kim R. Robertson and James G.
Helgeson (9186) “Leveraged” Decision Making in Advertising: The Flat Maximum Principle
and Its Implications, Journal of Marketing Research, Vol. 23, No. 1 (Feb.), 25-32.
324
MARKETING METRICS
10
ONLINE, EMAIL, AND
MOBILE METRICS
Introduction
Key concepts covered in this chapter:
Impressions, Pageviews, and Hits
Visits, Visitors, and Abandonment
Rich Media Display Time
Bounce Rate
Rich Media Interaction Rate
Friends/Followers/Supporters, Likes
Clickthrough Rates
Downloads
Cost per Impression, Cost per Click,
and Cost of Acquisition
Email Metrics
Mobile Metrics
In this chapter, we focus on metrics used in Web-based and other marketing efforts
enabled by the widespread use of information technology. The Internet increasingly
provides valuable opportunities to augment traditional broadcast and print advertising
with interactive media. In fact, many advertising media terms, such as impressions, are
used to describe and evaluate Web-based advertising while other terms, such as clickthrough, are unique to the Web. Certain Web-specific metrics are needed because the
Internet, like direct mail, serves not only as a communications medium, but also as a
direct sales channel that can provide real-time feedback on the effectiveness of advertising in generating customer interest and sales.
The potential of the Internet for marketers is easy to see. With respect to metrics,
the Internet provides an extremely rich source of data. Indeed marketers working in
this area often have a profusion of data sources unlike many in traditional areas of
325
marketing. Lots of data is only useful if marketers know what to do with it. This chapter
outlines some key metrics and the strengths and weaknesses of these metrics.
Metric
Construction
Considerations
Purpose
10.1
Pageviews
The number of
times a Web page is
served.
Represents the
number of Web
pages served.
To provide a toplevel measure of the
popularity of a Web
site.
10.2
Rich Media
Display Time
The average time
that rich media are
displayed per viewer.
Can be heavily
influenced by
unusually long
display times. How
data is gathered
is an important
consideration.
To measure average
viewing time of rich
media.
10.2
Rich Media
Interaction
Rate
Provides fraction of
viewers interacting
with the rich media.
The definition of
interaction should
exclude actions
unrelated to the rich
media (a mouse
crossing the rich
media to reach
another part of the
screen).
Measures relative
attractiveness of
rich media and
ability to generate
viewer engagement.
10.3
Clickthrough
Rate
Number of
clickthroughs
as a fraction of
the number of
impressions.
An interactive
measure of Web
advertising. Has
great strengths, but
clicks represent
only a step toward
conversion and are
thus an intermediate
advertising goal.
To measure the
effectiveness of a
Web advertisement
by counting those
customers who
are sufficiently
intrigued to click
through it.
10.4
Cost per Click Advertising
cost, divided by
number of clicks
generated.
Often used as a
billing mechanism.
To measure or
establish the cost
effectiveness of
advertising.
326
MARKETING METRICS
Metric
Construction
Considerations
10.4
Cost per
Order
Advertising cost,
More directly
divided by number of related to profit
orders generated.
than cost per click,
but less effective
in measuring the
impact of pure
advertising. An
advertisement may
generate strong
clickthrough
but yield weak
conversion due to
a disappointing
product.
To measure or
establish the costeffectiveness of
advertising.
10.4
Cost per
Impression
Advertising cost,
Useful for purposes
divided by number of of comparison to
customers acquired.
customer lifetime
value. Helps
marketers determine
whether customers
are worth the cost of
their acquisition.
To measure or
establish the costeffectiveness of
advertising.
10.5
Visits
The number of times By measuring
that visitors come to a visits relative
Web site.
to pageviews,
marketers can
determine
whether viewers
are investigating
multiple pages on a
Web site.
To measure
audience traffic on
a Web site.
10.5
Visitors
The number of
unique Web site
visitors in a given
period.
To measure the
reach of a Web site.
Useful in
determining the
type of traffic
generated by a Web
site—a few loyal
adherents, or many
occasional visitors.
The period over
which this metric
is measured can
be an important
consideration.
Purpose
Chapter 10 Online, Email, and Mobile Metrics
327
Metric
Construction
Considerations
Purpose
10.5
Abandonment The rate of purchases Can warn of
Rate
started but not
weak design in an
completed.
e-commerce site
by measuring the
number of potential
customers who
lose patience with a
transaction process
or are surprised
and put off by
“hidden” costs
revealed toward its
conclusion.
To measure one
element of the close
rate of an Internet
business.
10.6
Bounce Rate
(Web site)
Fraction of Web site Requires a clear
visitors who only view definition of when
a single page.
a visit ends. Usually
considers bounce
rate with respect to
visits rather than
visitors.
Often used as an
indicator of site’s
relevance and
ability to generate
visitor interest.
10.7
Friends/
Followers/
Supporters
Number of
individuals joining a
social network.
Success depends on
target group and
the social nature of
the product. This
metric is unlikely to
reflect the ultimate
aim of a marketing
campaign.
To measure size of
social network, but
unlikely to measure
engagement.
Likes
Number who have
favored a post/site/
organization on social
media.
Measures
engagement with
the liked entity.
Can be used to
understand what
posts resonate more
with consumers.
Understand the
relative popularity
of social media
elements.
10.7
328
MARKETING METRICS
Metric
Construction
Considerations
Purpose
10.7
Value of a
Like
A dollar value
ascribed to each like
(or follower etc.).
Beware of causal
claims. Consumers
usually like a
product on a social
network because
they favor the
product offline.
Merely generating
likes won’t
necessarily change
purchase behavior.
To give a dollar
value for like.
(Do not use as a
benchmark for
spending on social
media.)
10.8
Downloads
Number of times an
application or file is
downloaded.
Counts the times a
file was downloaded,
not the number
of customers who
downloaded a file.
It is often useful to
monitor downloads
started but not
completed.
To determine
effectiveness in
getting applications
out to users.
10.9
Mobile
Metrics
Revenues divided by
the number of users.
May help to indicate
how effective an
app is at generating
revenues but
cost figures are
needed to establish
profitability.
Examines
effectiveness of
app at generating
revenue from its
user base.
10.10
Email
Clickthrough
Clickthrough rate is
Measures how
percentage of times
effective email is at
an email is clicked on. gaining attention.
To determine
effectiveness of
email campaigns.
10.1 Impressions, Pageviews, and Hits
As noted in Section 9.1, impressions represent the number of opportunities that
have been presented to people to see an advertisement. The best available measures
of this figure use technology in an effort to judge whether a given advertisement was
actually seen. But this is never perfect. Many recorded impressions are not actually
Chapter 10 Online, Email, and Mobile Metrics
329
perceived by the intended viewer. Consequently, some marketers refer to this metric
as opportunities-to-see.
In applying this concept to Internet advertising and publishing, pageviews represent the number of opportunities-to-see for a given Web page. Every Web page is
composed of a variety of individual objects and files, which can contain text, images,
audio, and video. The total number of these files requested in a given period is the
number of hits a Web site or Web server receives. Measuring hits has largely been
abandoned because pages composed of many small files generate numerous hits per
pageview giving an overly positive view of consumer activity.
Purpose: To assess Web site traffic and activity.
To quantify the traffic a Web site generates, marketers monitor pageviews—the number
of times a page on a Web site is accessed. In the early days of e-commerce, managers
paid attention to the number of hits a Web site received. Hits measure file requests.
Because Web pages are composed of numerous text, graphic, and multimedia files, the
hits pages receive are a function not only of pageviews, but also of the way those pages
were composed by their Web designer. As marketing on the Internet has become more
sophisticated, better measures of Web activity and traffic have evolved. As hits can be
influenced by Web page design, this measure is unhelpful and old fashioned and it is
now much more common to use pageviews to measure traffic at a Web location.
Pageviews aim to measure the number of times a page has been displayed to a user. It
should be measured as close to the end user as possible. The best technology counts
pixels returned to a server, confirming that a page was properly displayed. This pixel1
count technique yields numbers closer to the end user than would a tabulation of
requests to the server, or of pages sent from the server in response to a request. Good
measurement can mitigate the problems of inflated counts due to servers not acting
on requests, files failing to serve on a user’s machine, or users terminating the serving
of ads.
Pageviews: The number of times a specific page has been displayed to users. This
should be recorded as late in the page-delivery process as possible in order to get as
close as possible to the user’s opportunity to see. A page can be composed of multiple
files.
A further distinction needs to be made as to how many times an advertisement was
viewed by unique visitors. A marketer is typically interested in how many people
viewed an advertisement and the technology to serve different ads to different visitors is widely available. For example, two individuals entering a Web page from two
different countries might receive the page in their respective languages and will also
330
MARKETING METRICS
probably receive different advertisements. One example of an advertisement that is
commonly displayed differently to different visitors is an embedded link with a banner ad. Recognizing this potential for variation, advertisers want to know the number
of times that their specific advertisement was displayed to visitors rather than just the
site’s number of pageviews.
With this in mind, Internet advertisers often perform their analyses in terms of impressions—sometimes called ad impressions or ad views. These represent the number
of times an advertisement is served to visitors, giving them opportunities to see it.
(Many of the concepts in this section are in line with the terms covered in Chapter 9,
“Advertising Metrics.”) Of course, if a page carries multiple advertisements, the total
number of all ad impressions will exceed the number of pageviews.
Major Web sites will have an ad server involved in presenting advertisements. The ad
server can control who sees what and may serve different ads to different visitors or
multiple ads to a single visitor. This can get surprisingly complex and technical issues,
such as the page refreshing, may complicate the data. Given the variety of ways the
advertisement can be seen, it can be a challenge to reconcile the data between the Web
site and ad server.
Construction
Hits: A count of the number of files served to visitors on the Web. Because Web pages
often contain multiple files, hits is a function not only of pages visited, but also of the
number of files on each page. While we detail this metric given its major historical significance, hit counts are likely to be more relevant to technicians responsible for planning server capacity than to marketers interested in measuring visitor activity.
Hits (#) = Pageviews (#) * Files on the Page (#)
Example: If a Web site served three files per page and generated 300,000 pageviews,
then hits would total 3 * 300,000 = 900,000.
Pageviews: The number of pageviews can be easily calculated by dividing the number
of hits by the number of files on the page.
Pageviews (#) =
Hits (#)
Files on Page (#)
Example: There are 250,000 hits on a Web site that serves five files each time a page is
accessed. Pageviews = 250,000/5 = 50,000.
Chapter 10 Online, Email, and Mobile Metrics
331
Data Sources, Complications, and Cautions
Pageviews, page impressions, and ad impressions are measures of the responses of a
Web server to page and ad requests from users’ browsers, filtered to remove robotic
activity and error codes prior to reporting. These measures are recorded as close as
possible to the user’s opportunity to see the page or ad to most accurately capture the
opportunities.2
For very simple Web sites a count of ad impressions can be derived from pageviews
if the percentage of pageviews that contain the ad in question is known. For example,
if 10% of pageviews receive the advertisement for a luxury car, then the impressions
for that car ad will equal 10% of pageviews. Such simple Web sites that serve the same
advertisement to all Web users are easy to monitor, but for most marketers working in
major firms life is more challenging. They typically work with ad servers that will have
their own reporting systems to cope with the measurement problems.
The impression-based metrics quantify opportunities-to-see; they do not take into
account the number of ads actually seen or the quality of what is shown. In particular,
these metrics do not account for the following:
■
Whether the message appeared to a specific, relevant, defined audience.
■
Whether the people to whom the pages appeared actually looked at them.
■
■
Whether the advertisement was clearly visible. The term “below the fold,” borrowed from newspapers, is used to describe advertisements on a page that are
not visible in the initial display of the page. Generally, if visitors must scroll to
see the advertisement they are less likely to be influenced by it.
Whether those who looked at the page had any recall of the page’s content, or
of any advertising messages contained on it, after the event.
Despite the use of the term “impression,” these measures do not tell a business manager
about the effect that an advertisement has on potential customers, i.e., whether the
advertisement actually made an impression on the potential customer. Marketers can’t
be sure of the effect that pageviews have on visitors.
Finally, often pageview results will consist of data that include duplicate showings to
the same visitor. For this reason, the term gross impressions may be used instead to
highlight what is often a key point—that opportunities-to-see may be delivered to the
same viewer on multiple occasions.
332
MARKETING METRICS
10.2 Rich Media Display Time and Interaction Rate
Marketers use the rich media display time metric to monitor how long their advertisements are holding the attention of potential customers.
Average Rich Media Display Time (#) =
Total Rich Media Display Time (#)
Total Rich Media Impressions (#)
Rich media display time represents an important way of tracking the success of
Internet advertising.
Marketers use the rich media interaction rate to assess the effectiveness of a single rich
media advertisement in generating engagement from its viewers.
Rich Media Interaction Rate (%) =
Total Rich Media Impressions with Interactions (#)
Total Rich Media Impressions (#)
Rich media interaction rate represents an important way of tracking the success of
Internet advertising in that it monitors the fraction of impressions that generate interaction on the part of the viewer.
Purpose: To determine how an advertisement engages viewers
Rich media is a term used for interactive media that allow consumers to be more actively
engaged than they might be with a billboard, a TV advertisement, or even a static display Web advertisement. Rich media metrics, or Audience Interaction Metrics, are very
similar in principle to other advertising metrics. Marketers want to track whether the
advertisement is effective at grabbing and maintaining the attention of potential customers and so they track how long people spend “viewing” the advertisement as a proxy
for how interested they are in the content of the advertisement. The rich media display
time shows how long, on average, people spend engaged with the rich media.
The rich media interaction rate tracks how actively involved potential consumers are
with an advertisement. The big advantage of rich media is the ability of viewers to
interact with it. Marketers using rich media can have a much better idea of potential
customers’ reactions to an advertisement simply because these interactions are counted.
They can monitor whether potential customers are simply passively “viewing” the
media on their screen or are actively engaged by taking some traceable action. A user
who interacts is showing evidence of being more actively engaged and is thus probably
more likely to move toward purchase.
Chapter 10 Online, Email, and Mobile Metrics
333
Construction
Average Rich Media Display Time: Rich media display time is simply the average time
that viewers spent with the rich media of an advertisement. For this the marketer will
need the total amount of time spent with the rich media and the total number of times
that the rich media was displayed. It is a simple matter to create an average time in seconds spent with the rich media by dividing the total amount of time in seconds spent
by the total number of impressions.
Average Rich Media Display Time (#) =
Total Rich Media Display Time (#)
Total Rich Media Impressions (#)
Rich Media Interaction Rate: This metric is the number of impressions of an advertisement that were interacted with divided by the total number of impressions of that
advertisement. It tells the marketers how successful any advertisement was at getting
potential customers to engage with it in some way, (mouse rollover, click on, deliberately starting a video, etc.). As an example, a rich media advertisement that was displayed 100 times with an interaction rate of 15% would mean that 15 of the impressions
resulted in some kind of interaction whereas 85 resulted in no interaction.
Rich Media Interaction Rate (%) =
Total Rich Media Impressions with Interactions (#)
Total Rich Media Impressions (#)
Data Sources, Complications, and Cautions
As with many Web-based metrics, data often seem abundant to marketers who come
from the offline world. However, there are several measurement issues the marketer
must address in order to convert the abundance of data into useful metrics (i.e., to
information and eventually to knowledge). For example, marketers usually cut display
times off at some upper bound; that is, if the piece of rich media has been displayed for
five minutes, it is safe to assume the viewer has probably gone to make a cup of coffee or
been otherwise distracted. The question of how long a displayed piece of rich media was
actually viewed has some similarities to the question offline marketers face with respect
to whether an offline advertisement was viewed. A slight advantage here goes to the rich
online media in that most displays of rich media begin because of an active request of
the viewer…whereas no such action is required offline.
Rich media display time, because it usually deals with short periods of time, can be
influenced by unusual events. Take a simplified example: If five people see the rich
media display for one second each and one person sees it for 55 seconds, the (average) rich media display time is ten seconds. There is no way to distinguish this average
334
MARKETING METRICS
display time from the average time generated by six moderately interested viewers each
viewing the advertisement for ten seconds. Such is the case with any average.
Marketers should be clear that they understand how the data were gathered and be
especially aware of any changes in the way the data were gathered. Changes in the way
the data were gathered and the metric constructed may be necessary for technological
reasons, but will limit the usefulness of the metric as longitudinal comparisons, i.e.,
over time, are no longer valid. At a minimum, the marketer must be aware of and
account for measurement changes when interpreting the metric.
Data for the rich media interaction rate will typically be available. Indeed the metric
itself might be reported as part of a standard reporting package. One important decision that has to be made in generating the metric is what counts as an interaction.
This will depend upon the potential actions that the viewers could take, which in turn
depends upon the precise form of the advertisement. What counts as an interaction will
usually have some lower bound. For example, an interaction is only counted if the visitor spends more than one second with his mouse over the impression. (This is designed
to exclude movements of the mouse unrelated to the advertisement such as moving the
mouse over the advertisement to another part of the page.)
As is true of any advertising, marketers should not forget the goal of their advertising.
Interaction is unlikely to be an end in itself. As such, a larger interaction rate, which
might be secured by gimmicks that appeal to people who will never buy the product,
may be no better than a smaller rate if the larger rate doesn’t move the visitor closer to
a sale (or some other high order objective).
Related Metrics
Rich Media Interaction Time: This metric captures the total amount of time that a visitor spends interacting with an advertisement. This is an accumulation of the total time
spent interacting per visit on a single page. On a visit to a page a user might interact
with the rich media for two interactions of two seconds each, and so have an interaction
time of four seconds.
Video Interactions: Video metrics are very similar to rich media metrics. Indeed
video can be classified as rich media depending upon the way it is served to the viewer.
Similar principles apply, and the marketer should track how long viewers engage with
the video (the amount of time the video plays), what viewers do with the video (pause
it, mute it), and the total and specific interactions with the video (which show evidence
of attention to the video). Such metrics are then summarized across the entire pool of
visitors, (for instance the average visit led to the video being played for 12 seconds).
Chapter 10 Online, Email, and Mobile Metrics
335
10.3 Clickthrough Rates
Clickthrough rate is the percentage of impressions that lead a user to click on an ad.
It describes the fraction of impressions that motivate users to click on a link, causing
a redirect to another Web location.
Clickthrough Rate (%) =
Clickthroughs (#)
Impressions (#)
Most Internet-based businesses use clickthrough metrics. Although these metrics are
useful, they should not dominate all marketing analysis. Unless a user clicks on a “Buy
Now” button, clickthroughs measure only one step along the path toward a final sale.
Purpose: To capture customers’ initial response to Web sites.
Most commercial Web sites are designed to elicit some sort of action, whether it be to
buy a book, read a news article, watch a music video, or search for a flight. People generally don’t visit a Web site with the intention of viewing advertisements, just as people
rarely watch TV with the purpose of consuming commercials. As marketers, we want to
know the reaction of the Web visitor. Under current technology, it is nearly impossible
to fully quantify the emotional reaction to the site and the effect of that site on the firm’s
brand. One piece of information that is easy to acquire, however, is the clickthrough
rate. The clickthrough rate measures the proportion of impressions that led to an initiated action with respect to an advertisement that redirected the visitor to another page
where they might purchase an item or learn more about a product or service. Here we
have used “clicked their mouse” on the advertisement (or link) because this is the generally used term, although other interactions are possible.
Construction
Clickthrough Rate: The clickthrough rate is the number of times a click is made on the
advertisement divided by the total impressions (the times an advertisement was served).
Clickthrough Rate (%) =
Clickthroughs (#)
Impressions (#)
Clickthroughs: If you have the clickthrough rate and the number of impressions, you
can calculate the absolute number of clickthroughs by multiplying the clickthrough rate
by the impressions.
Clickthroughs (#) = Clickthrough Rate (%) * Impressions (#)
336
MARKETING METRICS
Example: There are 1,000 clicks (the more commonly used shorthand for clickthroughs) on a Web site that serves up 100,000 impressions. The clickthrough rate
is 1%.
Clickthrough Rate (%) =
1,000
100,000
= 1%
If the same Web site had a clickthrough rate of 0.5%, then there would have been 500
clicks:
Clickthrough Rate = 100,000 * 0.5% = 500
If a different Web site had a 1% clickthrough rate and served up 200,000 impressions,
there would have been 2,000 clicks:
# of Clicks = 1% * 200,000 = 2,000
Data Sources, Complications, and Cautions
The number of impressions is a necessary input for the calculation. On the very simplest
Web sites, this is likely to be the same as pageviews; every time the page is accessed, it
shows the same details. On more sophisticated sites, different advertisements can be
shown to different viewers meaning the number of impressions of the advertisement
cannot simply be estimated from the number of pageviews. The good news is that this
metric is usually available as part of a reporting package. For clickthrough rate you also
need clicks, and the server can easily record the number of times the link was clicked
(see Figure 10.1).
Clickthrough rate
captures numbers
clicking on ad.
New Page
Served
Clicked
Potential
Customer
Cancelled Before Ad Served
Didn’t Click
Out of Process
Figure 10.1
Clickthrough Process
Chapter 10 Online, Email, and Mobile Metrics
337
First, remember that clickthrough rate is expressed as a percentage. Although high
clickthrough rates might in themselves be desirable and help validate your ad’s appeal,
companies will also be interested in the total number of people who clicked through.
Imagine a Web site with a clickthrough rate of 80%. It may seem like a highly successful Web site until management uncovers that only 20 people visited the site. Rather
than the healthy 80% clickthrough rate, the resulting 16 potential customers who
clicked through might be better compared with management’s stated objective of 500
clickthroughs.
Also remember that a click is a very weak signal of interest. Individuals who click on an
ad might move on to something else even before the new page is loaded. This could be
because the person clicked on the advertisement by accident or because the page took
too long to load. This is a problem that is of greater significance with the increase in
richer media advertisements. Marketers should understand whether their customers
are using devices able to quickly display the requested file. Using large video files is
likely to increase the number of people abandoning the process before the ad is served,
especially if the potential customers have slower connections.
As with impressions, try to ensure that you understand the measures. If the measure is
of clicks (the requests received from client machines to the server to send a file), then
there may be a number of breakage points between the click and the impressions of the
ad generated from a returned pixel count. Large discrepancies should be understood—
what role is played by technical problems (the size/design of the advertisement) and
what is the role of weak interest from clickers?
Clicks are the number of times the advertisement was interacted with, not the number
of customers who clicked. An individual visitor can click on an ad several times—
either in a single session or across multiple sessions. You will have to investigate your
Web analytics package to get a deeper understanding such as how many times an ad
was clicked on by someone using the same browser. Furthermore, sophisticated Web
sites can control the number of times they show a specific advertisement to the same
customer. (A customer who logs onto a site can be tracked which gives more information on what they see.) Finally, the clickthrough rate must be interpreted relative to
an appropriate baseline. Clickthrough rates for banner ads are very low and continue
to fall. In contrast, clickthrough rates for buttons that simply take visitors to the next
page on a site should be much higher. An analysis of how clickthrough rates change as
visitors navigate through various pages can help identify “dead end” pages that visitors
rarely move beyond.
One can test the effectiveness of various Web page designs by serving different pages
randomly. The random serving of pages, when done correctly, allows the marketer to
assume that those viewing the page are similar and any differences in reaction can thus
be attributed to differences in the Web design.
338
MARKETING METRICS
It is possible to test a range of different sites and variations of content on the sites easily. These are commonly known as A/B tests, or multivariate tests when there is more
than one difference. There are now many tools available to help run these tests; e.g.,
Unbounce, that are affordable for a wide variety of organizations. In general, best practice is to test to help find the best version of a Web page or advertisement possible.
One useful approach is to classify metrics such as clickthroughs as micro-conversions
on the path to the macro-conversion, an eventual purchase. The aim of the firm is not
to generate clickthroughs, the micro-conversions, per se but the micro-conversions are
a necessary stage in attaining the final aim, the macro-conversion. Describing them
as macro- and micro-conversions helps keep the focus on the macro-conversion—the
ultimate goal—while still realizing that failing on micro goals tends to mean you will
ultimately fail on the macro goal.
10.4 Cost per Impression, Cost per Click, and Cost per Order
These metrics measure the average cost of impressions, clicks, and customers. All are
calculated in the same way—as the ratio of cost to the number of resulting impressions, clicks, or customers.
Cost per Impression ($) =
CPM ($) =
Advertising Cost ($)
Number of Impressions (#)
Advertising Cost ($)
Number of Impressions in Thousands (#)
Cost per Click ($) =
Cost per Order ($) =
Advertising Cost ($)
Number of Clicks (#)
Advertising Cost ($)
Orders (#)
These metrics are the starting point for assessing the effectiveness of a company’s
Internet advertising and can be used for comparison across advertising media and
vehicles.
Chapter 10 Online, Email, and Mobile Metrics
339
Purpose: To assess the cost effectiveness of Internet marketing.
In this section, we present three common ways of measuring the cost effectiveness of
Internet advertising. Each has benefits depending upon the perspective and end goal of
the advertising activity.
Cost per Impression gives the cost to offer potential customers one opportunity to see
an advertisement. CPM is the more commonly used version of cost per impression. It
is the same metric but using per thousand impressions as the denominator. (For some
reason hi-tech people like to use mille, the Latin for thousand, hence the M in CPM.)
Cost per Click gives the average amount spent to get an advertisement clicked. Cost per
click has a big advantage over cost per impression in that it tells us something about
how effective the advertising was. Clicks are a way to measure attention and interest.
An advertisement may look inexpensive in that it has a relatively low CPM. Of course,
if the advertisement isn’t relevant it will end up with few clicks and so a high cost per
click. Badly targeted adverts can be very expensive as measured by Cost per Click. If the
main purpose of an ad is to generate a click, then cost per click is the preferred metric.
Cost per Order gives the average cost to acquire an order. If the main purpose of the ad
is to generate sales, then cost per order is the preferred metric.
A certain number of Web impressions need to be achieved to generate a reasonable
number of orders and the quality and placement of the advertisement will affect both
click(through) rates and the resulting cost per click (see Figure 10.2).
Further along, measures are better tied to overall business objectives.
Earlier in the process, measures are less affected by noise.
Potential
Customer
Doesn’t See Ad
Order
Placed
Cost per
Order
Follows
Link
Cost per
Click
Sees Ad
Cost per
Impressions
Doesn’t Click
Doesn’t Buy
Customer Out of Process
Figure 10.2
340
MARKETING METRICS
The Order Acquisition Process
Construction
The formulas are essentially the same for the alternatives; just divide the advertising
cost by the appropriate number; for example, impressions, clicks, or orders.
Cost per Impression: This is derived from advertising cost and the number of
impressions.
Cost per Impression ($) =
Advertising Cost ($)
Number of Impressions (#)
Cost per impression is typically expressed as cost per thousand impressions (CPM) in
order to make the numbers easier to manage (for more on CPM, refer to Section 9.2).
Cost per Click: This is calculated by dividing the advertising cost by the number of
clicks generated by the advertisement.
Cost per Click ($) =
Advertising Cost ($)
Number of Clicks (#)
Cost per Order: This is the cost to generate an order. The precise form of this cost
depends on the industry and is complicated by product returns and multiple sales channels. That said the basic formula is
Cost per Order ($) =
Advertising Cost ($)
Orders (#)
EXAMPLE: An Internet retailer spent $24,000 on online advertising and generated 1.2 million impressions.
$24,000
Cost per Impression ($) =
= $0.02
1,200,000
This led to 20,000 clicks.
Cost per Click ($) =
$24,000
20,000
= $1.20
Only 1 in 10 of the clicks resulted in a purchase.
Cost per Order ($) =
$24,000
2,000
= $12.00
This last calculation is also called “cost per purchase.”
Cost per Customer Acquired: Simply divide the advertising cost by the number of new
customers who make an order. Refer to Chapter 5, “Customer Profitability,” for more
on defining customer and acquisition costs.
Chapter 10 Online, Email, and Mobile Metrics
341
Data Sources, Complications, and Cautions
The Internet has provided greater availability of advertising data. Consequently,
Internet advertising metrics are likely to rely on data that is more readily obtainable
than data from conventional channels. The Internet can provide more information
about how customers move through the system and how individual customers behave
at the purchase stage of the process.
The calculations and data we have discussed in this section are often used in contracts
compensating advertisers. Companies may prefer to compensate media and ad agencies on the basis of new customers acquired instead of impressions although the agencies may be less happy as it pushes more risk onto the agency. The agency may find
its efforts limited by the lack of appeal of the client’s products rather than the agency’s
efforts to produce interesting advertising and get it widely distributed.
Search Engines
Search engine payments help determine the placement of links on search results pages.
An important search engine metric is the cost per click, and it is generally the basis for
establishing the search engine placement fee. Search engines can provide plenty of data
to analyze the effectiveness of a campaign. In order to reap the benefits of a great Web
site, the firm needs to get people to visit it. Previously we discussed how firms measure
traffic. Search engines help firms create that traffic.
Although a strong brand helps drive traffic to a firm’s site, including the firm’s Web
address in all of its offline advertising might not be sufficient to increase traffic count.
In order to generate additional traffic, firms often turn to search engines. EMarketer
estimated that social network spending reached $16.1 billion in 2014. Furthermore,
overall $179 billion was spent on paid media in 2014 in the USA. Digital spending
accounted for a significant 28.2% of this total.3
Paid search marketing (also known as search engine marketing) is essentially paying
for the placement of ads on search engines across the Internet. The ads are typically
small portions of text (much like newspaper want ads) made to look like the results of
an unpaid or organic search. Payment is usually made only when someone clicks on
the ad. It is possible to pay more per click in return for better placement on the search
results page. Advertisers can bid to be displayed whenever someone searches for specified keyword(s). In this case, companies bid on the basis of cost per click. Bidding a
higher amount per click gets you placed higher. However, there is an added complexity,
which is if the ad fails to generate several clicks, its placement will be lowered in comparison to competing ads despite a higher CPC bid.
The measures for testing search engine effectiveness are largely the same as those used
in assessing other Internet advertising.
342
MARKETING METRICS
Cost per Click: The critical concept in search engine marketing is cost per click. Cost
per click is widely quoted and used by search engines in charging for their services.
Marketers use cost per click to build their budgets for search engine payments.
Search engines ask for a “maximum cost per click,” which is a ceiling whereby the marketer imposes the maximum amount they are willing to pay for an individual click. A
search engine will typically auction the placement of links and only charge for a click
at a rate just above the next highest bid. This means the maximum cost per click that a
company would be willing to pay can be considerably higher than the average cost per
click they end up paying.
Marketers often talk about the concept of daily spend on search engines—just as it
sounds, this is the total spent on paid search engine advertising during one day. In
order to control spending, search engines allow marketers to specify maximum daily
spends. When the maximum is reached, the advertisement doesn’t show again until the
next 24 hour period.
Daily spend can be thought of as the product of average cost per click and the number
of clicks:
Daily Spend ($) = Average Cost per Click ($) * Number of Clicks (#)
EXAMPLE: Andrei, the Internet marketing manager of an online music retailer, decides to set
a maximum price of $0.10 a click. At the end of the week he finds that the search engine provider
has charged him a total of $350.00 for 1,000 clicks per day.
His average cost per click is thus the cost of the advertising divided by the number of clicks
generated:
Cost per Week
Cost per Click ($) =
Clicks per Week
=
$350
7,000
=$0.05 a Click
His average daily spend is $350/7 or $50 per day. This is average daily spend, which is not to be
confused with daily spend limit.
Ad Rank: This is the position of an advertisement served on a Pay Per click basis (PPC)
on a search engine. The rank depends upon the amount bid for each keyword as well as
the relevance of the keywords, which determines the quality and ranking.
Chapter 10 Online, Email, and Mobile Metrics
343
Search engines typically use auctions to establish a price for the search terms they sell
and have the great advantage of having a relatively efficient market; all users have access
to the information and can be in the same virtual location. They tend to adopt a variant
on the second price auction. Buyers only pay the amount needed for their requested
placement; as such, search engine marketers can control the price they are willing to
pay. The trick therefore is to know how much is reasonable for your firm to pay per
click, which ultimately is a managerial judgment depending upon the benefits you
expect to receive from such things as the customers generated.
Search Engine Optimization
Search engine optimization involves efforts to make your Web site rank more highly
on search engines’ organic (unpaid) search results. There is little reason not to try
and improve your ranking on organic search. Of course in practice successful search
engine optimization involves considerable skill and knowledge of how the rankings are
constructed. The search engines have a variety of algorithms for how they rank sites.
Possibly the most famous is Google’s PageRank, which works by counting the number
and quality of links to a page to determine a rough estimate of how important the Web
site is.
10.5 Visits, Visitors, and Abandonment
Visits measures the number of sessions on the Web site. Visitors measures the number of people making those visits. When an individual goes to a Web site on Tuesday
and then again on Wednesday, this should be recorded as two visits from one visitor in the week. Visitors are sometimes referred to as “unique visitors.” Visitors and
unique visitors are the same metric.
Abandonment usually refers to shopping carts. The total number of shopping carts
used in a specified period is the sum of the number abandoned and the number that
resulted in complete purchases. The abandonment rate is the ratio of the number of
abandoned shopping carts to the total.
Purpose: To understand Web site user behavior.
Web sites can easily track the number of pages requested. As we saw earlier in Section
10.1, the pageviews metric can be useful but is far from a complete metric. We can do
better. In addition to counting the number of pageviews a Web site delivers, firms will
also want to count the number of times someone visits the Web site and the number of
people requesting those pages.
344
MARKETING METRICS
To get a better understanding of traffic on a Web site, companies attempt to track the
number of visits. A visit (known as a “session” in Google Analytics) can consist of a
single pageview or multiple pageviews, and one individual can make multiple visits to a
Web site. Visits captures the number of times individuals request a page from a website
for the first time, i.e., the first request triggers the visit. Subsequent requests from the
same individual do not count as visits unless they occur after a specified timeout period.
The exact specification of what constitutes a visit requires this accepted standard for a
timeout period, usually set at 30 minutes, which is the number of minutes of inactivity
from the time of entering the page to the time of requesting a new page.
In addition to visits, firms also attempt to track the number of individual visitors to
their Web site. Visitors captures the number of individuals requesting pages from the
firm’s Web site server during a given period; also known as unique visitors. Google
Analytics refers to these as “users.” Because a visitor can make multiple visits in a specified period, the number of visits will be greater than the number of visitors. A visitor is
sometimes referred to as a unique visitor or unique user to clearly convey the idea that
each visitor is only counted once.
The measurement of users or visitors requires it to be measured over a time period, e.g.,
number of visitors in a given month.
Pageviews and visits are related. By definition, a visit is a series of pageviews grouped
together in a single session, so the number of pageviews will exceed the number of visits.
Consider the metrics as a series of concentric ovals as shown in Figure 10.3. In this
view, the number of visitors must be less than or equal to the number of visits, which
must be less than or equal to the number of pageviews, which must be equal to or less
than the number of hits. (Refer to Section 10.1 for details of the relationship between
hits and pageviews.)
Another way to consider the relationship between visitors, visits, pageviews, and hits is
to consider the following example of one visitor entering a Web site of an online newspaper (see Figure 10.4). Suppose that the visitor enters the site on Monday, Tuesday,
and Friday. In her visits she looks at a total of 20 pageviews. Those pages are made up
of a number of different graphic files, word files, and banner ads.
The ratio of pageviews to visitors is sometimes referred to as the average pages per visit.
Marketers track this average to monitor how the average visit length changes over time.
It is possible to dig even deeper and track the paths visitors take. This path is called the
clickstream.
Clickstream: The path of a user through the Internet.
Chapter 10 Online, Email, and Mobile Metrics
345
Hits
Pageviews
Visits
Visitors
Figure 10.3
1 Visitor
Relationship of Hits to Pageviews to Visits to Visitors
3 Visits
Monday
Tuesday
Friday
20 Pageviews
200 Hits
10 News
5 Sports
5 Business
160 graphic files
20 word files
20 banner ads
Figure 10.4 Example of Online Newspaper Visitor
The clickstream refers to the sequence of clicked links the user makes. Using the clickstream on his or her own site a marketer tracking at this level can help the firm identify
the most and least appealing pages (see Figure 10.5).
The analysis of clickstream data often yields significant customer insights. What path
is a customer most likely to take prior to purchase? Is there a way to make the most
popular paths even easier to navigate? Should the unpopular paths be changed or even
eliminated? Do purchases come at the end of lengthy or short sessions? At what pages
do sessions end?
346
MARKETING METRICS
News
In Depth
Sales
Welcome
Features
Links not used
Member
Benefits
Member
Login
Clickstream, the actual path taken by a customer
Figure 10.5 A Clickstream Documented
A portion of the clickstream that deserves considerable attention is the subset of clicks
associated with the use of shopping carts. A shopping cart is a piece of software on the
server that allows visitors to select items for eventual purchase. Although shoppers in
brick and mortar stores rarely abandon their carts, abandonment of virtual shopping
carts is quite common. Savvy marketers count how many of the shopping carts used in
a specified period result in a completed sale versus how many are abandoned. The ratio
of the number of abandoned shopping carts to the total is the abandonment rate.
Construction
Visits and Visitors: Your analytics system will probably report these figure. Cookies
can help servers track unique visitors, but this data is never 100% accurate (see the next
section).
Abandonment Rate: This measures the percentage of carts that were abandoned before
completion. The percentage of shopping carts that are abandoned.
Abandonment Rate (%) =
Carts Not Completed (#)
Customer Baskets Initiated (#)
EXAMPLE: An online comics retailer found that of the 25,000 customers who loaded items
into their electronic baskets, only 20,000 actually purchased:
Purchases Not Completed = Purchases Initiated Less Purchases Completed
= 25,000-20,000 = 5,000
Chapter 10 Online, Email, and Mobile Metrics
347
Abandonment Rate (%) =
Carts Not Completed (#)
Customer Baskets Initiated (#)
=
5,000
25,000
= 20% abandonment rate
Conversion Rate: Many marketers monitor their conversion rate, the percentage of
visitors who actually buy. This gives a good headline view of how effective your Web
site is at generating sales once you have visitors.
Conversion Rate (%) =
Number of Consumers Who Purchase (#)
Number of Visitors (#)
Google Analytics uses sessions as the denominator. This means that each consumer can
convert more than once if they visit more than once. Be careful to understand which
your data represents.
Data Sources, Complications, and Cautions
Visits can be estimated from log file data. Visitors are much more difficult to measure.
Companies often encourage users to register so they can gain a better understanding of
their users. Often consumers aren’t keen to do so; perhaps because they see registering
as a hassle or as limiting their privacy. Indeed marketers rarely know for certain that a
user is unique. Instead they consider visits from new browsers evidence of a new user
and return visits from the same browser as a returning visitor. This means a user who
uses different devices will often show as two different users. Similarly different people,
for instance a family using a shared computer, may show as the same user.
To decide whether a visitor is a returning visitor or a new user, companies often employ
cookies. A cookie is a file downloaded onto the computer of a person surfing the Web
that contains identifying information. When the “browser” returns, the Web server
reads the cookie and recognizes the visitor as someone who has been to the Web site
previously. More advanced sites use cookies to offer customized content, and shopping
carts make use of cookies to distinguish one shopping cart from another. For example,
Amazon, eBay, and EasyJet all make extensive use of cookies to personalize the Web
views to each customer. As users become more sophisticated many are trying to protect
their anonymity by not allowing cookies. If visitors accept cookies, then at least the
browser that was used for the visit can be identified. If your customers do not accept
cookies, this is much harder and the marketer should consider how significant a problem this is for his or her operations.
348
MARKETING METRICS
Cookie: A small file that a Web site puts on the hard drive of visitors for the purpose
of future identification.
These metrics can be distorted by automatic activity (such as “bots”) that aim to classify
Web content. To prevent this, estimates of visitors, visits, and other traffic, statistics are
usually filtered to remove this activity by eliminating known IP addresses for “bots,” by
requiring registration or cookies, or by using panel data.
Consultants may be able to help quantify your activity. For example, Nielsen, among
other services, runs a panel in the U.S. and a number of major economies.4
10.6 Bounce Rate (Web Site)
Bounce Rate is a measure of the effectiveness of a Web site in encouraging visitors to
continue their visit. It is expressed as a percentage and represents the proportion of
visits that end on the first page of the Web site that the visitor sees.
Bounce Rate (%) =
Visits That Access Only A Single Page (#)
Total Visits To The Web Site (#)
High bounce rates typically indicate that the Web site isn’t doing a good job of attracting the continuing interest of visitors.
Purpose: To determine the effectiveness of the Web site at generating the
interest of visitors.
Bounce rate is a commonly reported metric that reflects the effectiveness of Web sites
at drawing the continuing attention of visitors. The assumption behind the usefulness
of the metric is that the owner of the Web site wants visitors to visit more than just the
landing page. For most sites this is a reasonable assumption. For example, sites that are
seeking to sell goods want visitors to go to other pages to view the goods and ultimately
make a purchase. Bounce rate is also a measure of how effective the company is at
generating relevant traffic. The more the Web site is relevant to the traffic coming to
it, the lower will be the bounce rate. This becomes particularly important when traffic
is generated through paid search. Money spent to generate traffic for whom the Web
site is not relevant (as reflected in a high bounce rate) is money wasted. The bounce
rate is a particularly useful measure with respect to the entry pages to Web sites. An
entry page with a very low bounce rate may be thought of as doing its job of driving
traffic to other pages but the actual rate depends upon a number of factors. As Google
Chapter 10 Online, Email, and Mobile Metrics
349
Analytics explains; “Evaluate and adjust factors that might contribute to your bounce
rate, like site layout and navigation. Use only your past performances as a rubric, and
try to improve your current bounce rate relative to your previous data. Provide enough
time between changes to collect enough data to evaluate the impact the changes may
be having on your users and their behaviors. Try using Content Experiments to help
you.”5 A content experiment, similar to an A/B test, is where you show different pages
to different visits (ideally randomly selected) and see which page performs best.
Having a low bounce rate is often a prerequisite of having a successful e-commerce
presence.
Construction
Bounce Rate: The number of visits that access only a single page of a Web site divided
by the total number of visits to the Web site.
Bounce Rate (%) =
Visits That Access Only A Single Page (#)
Total Visits To The Web Site (#)
Data Sources, Complications, and Cautions
Data to construct this metric, or even the metric itself, will usually come from the Web
site’s host as part of the normal reporting procedure. Given how common it is that
bounce rate is reported by default, it is a metric that is difficult to ignore. Construction
of the metric requires a precise definition of when a visit ends. Leaving the site may
come from closing the window, hitting the Back button or being timed out. (Google
Analytics counts session timeouts.) After a timeout a new session is usually started if
the visitor returns to the Web site. A lower timeout period results in increased bounce
rates, all else equal.
Reports may use the term visitors instead of visits. You should be clear what data is
actually reported. Visits are much easier to track because when the same visitor makes
return visits, especially to different entry pages, it can be difficult to connect the return
visit to the original visitor. As such visits, rather than visitors, are most likely used to
calculate bounce rates.
This metric can also be defined and constructed for individual landing pages within the
site. Indeed the bounce rate for each landing page allows for more precise diagnosis of
problem areas on a Web site. One must interpret page bounce rates, however, in light
of the purpose of the page. For some pages, such as directions pages where people go to
find out specific information, a high bounce rate is to be expected.
350
MARKETING METRICS
The value of the bounce rate metric will depend upon the objective of the organization.
Informational sites may develop a strong bond with their users through frequent short
interactions, such as when the users check sports scores. The organization may be comfortable if many users do not visit other parts of the site, and may not be too concerned
about high bounce rates. However, most companies will probably want low bounce
rates and will actively monitor this important metric.
One of the most challenging problems in marketing (online, offline, and online to
offline) is the problem of attribution. This is because a consumer is likely to have been
reached by multiple marketing actions prior to his or her purchase; making it impossible to say definitively which touchpoint generated the sales. Indeed all touchpoints
may have had a vital contribution in generating the sales. For advertisers using a mix
of online and “offline” media, it will be especially difficult to categorize the cause
and effect relationships between advertising and sales. Search ads might receive too
much credit for an order if the customer has also been influenced by the fact that
the consumer earlier saw the firm’s billboard advertisement. Conversely, search ads
might receive too little credit for offline sales where the consumer’s interest was previously piqued by the search ad. Similarly social media can sometimes play a big role in
sales; it assists the final conversion, but is usually not the last referral source clicked
before a purchase. Given this, social media doesn’t always get the credit it deserves for
driving sales.
Attribution is a challenge but various assumptions are made. When all items influencing a purchase are on-line, a firm can attribute the sale in a number of ways.
Last click attribution is when full credit for the sale is given to the final touchpoint prior
to purchase. First click attribution is when full credit for the sale is given to the initial
touchpoint, the first contact on the customer’s path to purchase. Linear attribution is
when the credit is shared equally between all touchpoints involved in the path to purchase. Clearly it is hard to defend any of these assumptions as perfectly capturing what
happens in the real-world but understanding the assumptions is necessary to determining whether the assumptions are useful approximations of reality even if they never
perfectly capture reality.
Google Analytics
This is not the place to give a detailed view of Google Analytics but the tools Google has
made available can be an excellent way of understanding the success of your marketing on a wide range of platforms. Google supplies an extensive range of assistance to
those wishing to use Google Analytics. This material, which includes graded courses to
work through, is supplied at the Google Analytics Academy, https://analyticsacademy.
withgoogle.com/explorer.
Chapter 10 Online, Email, and Mobile Metrics
351
10.7 Social Media Metrics: Friends/Followers/
Supporters/Likes
Friends/Followers/Supporters is a very simple metric that measures the number of
individuals who join an organization’s social network.
Friends (#) = Number of friends of the entity registered on a social network (#)
A high number of friends signifies an active interest in the owner of the page. If a
brand has a high number of friends, this indicates a stronger brand with a loyal customer base.
Likes is a similarly simple metric that measures the number of individuals who have
favored a post/page/organization using the “like” button.
Likes (#) = Number of individuals favoring a social networking post/page (#)
Purpose: To determine the effectiveness of a social networking presence.
We use the term friends to encompass followers, supporters, and other similar concepts.
Friends are members of a social networking site who register that they know, like, and/
or support the owner of the social networking page. For instance a strong brand may
have many customers who want to publicly signal their love of the brand. Social networking sites hold great benefits in allowing companies to develop customer relationships and can help a company identify and communicate with committed customers.
Likes are ways that individuals on social media show they favor a post or a page. They
click on the like button below pages, comments, brands, etc. Like is an extremely low
commitment activity, and some people like many things in a day. It seems, however,
a reasonable assumption that generally a larger number of likes is indicative of greater
appeal.
Construction
These metrics will be supplied by the social network.
Friends (#) = Number of friends of the entity registered on a social network (#)
Likes (#) = Number of individuals favoring a social networking post/page (#)
Cost per Friend (Like): The cost to the organization per friend recruited or Like
generated.
Cost per Friend ($) =
352
Total Costs to Provide Social Networking Presence ($)
MARKETING METRICS
Number of Friends (#)
Cost per Like ($) =
Total Costs to Provide Social Networking Presence ($)
Number of Likes (#)
Often the direct costs of having a social networking site are very low. This should
not, however, lead the marketer to conclude the cost is effectively zero. Sites have
to be designed, staff have to update the site, and marketers have to devise strategies.
Remember when calculating the cost of having a social network presence that the costs
should include all costs incurred in the provision of the social network presence.
Outcomes Per Friend: A similar attempt might be made to clarify the precise downstream outcomes gained by the presence of friends. (“Did we sell more ketchup?”) It is
often very hard to track and attribute outcomes to specific social networking actions.
This does not mean that an active social networking presence is not a vital part of an
Internet marketing strategy, but when designing a presence the ultimate objective of the
company needs to be borne in mind. For example, friends are often recruited to “vote”
in polls. The percentage of friends participating is a simple example of an “outcome per
friend” metric but probably not the ultimate objective.
Data Sources, Complications, and Cautions
Success in recruiting friends is likely to depend heavily on the group of people who
identify with the entity (e.g., individuals, brands, companies, or other groups). In the
case of brands, some customer segments are more reluctant to reveal their brand loyalty
than others, and as such two brands of equivalent strength may have very different levels of social network presence. Similarly the product involved is likely to influence the
likelihood of registering as a friend at the social networking site. It is easy to think of
some vitally important but more private products that are relied upon by their users but
are less likely to gain public expressions of support than brands that are more related to
public consumption.
It is very hard to objectively judge the effectiveness of social networking activities.
Generally having more followers or likes is an excellent sign of customer engagement.
The more customers who have an ongoing relationship with a brand that they are willing to publicly support, the more likely the brand is to have strong customer awareness
and loyalty. It is worth noting, however, that Friends and Likes, as with many metrics,
are most often intermediate metrics rather than an aim of the organization itself. It is
unlikely that most organizations exist with the explicit objective of generating Friends.
As such it is rarely sufficient to report the number of Friends as a successful outcome
of a marketing strategy without any additional information. This concern is only
increased by the widespread practice of selling “Friends” and “Followers” etc. on social
media. Likes need to be tied to the objectives of the firm and not just be a way of soothing the vanity of the marketer.
Chapter 10 Online, Email, and Mobile Metrics
353
It is often appropriate to construct metrics around the downstream outcomes and cost
effectiveness of such strategies. A marketer would be well advised to pay attention to the
costs and ultimate benefits of social networking presence as well as the clear potential
to engage with customers.
The idea of an Augmented Social Network (ASN) has become popular. ASN can be
used as a loose term to signify a social network that is successfully engaging with its
users. As such a measure, such as the percentage of completed social networks, or the
number of posts that are engaged with, may indicate successful engagement. A firm
might therefore measure the number of posts that generate a response divided by the
total number of posts to give some idea of engagement with the social network.
Value of a like is an attempt to put a dollar value on social media. It is possible to calculate the value of a like as the difference between the value of someone who doesn’t
like, for example, a brand and someone who does. Value of a like (value of a follow) is
sometimes calculated by subtracting the value of an individual consumer who did not
choose to like (or follow) a brand from the value of an individual who did. It is important to note that it would be wrong to argue that this difference is the value attributable
to the social media strategy. It is unlikely that the social media strategy caused the entire
difference in value observed between those choosing to like the brand and those not
doing so. There are likely a number of other factors.
There are, thus, two major caveats that we have seen with the use of this metric which
leads to concern that value of a like can cause more harm than good. Firstly, value to
a firm should be the profit generated by the consumer and not revenue. Assessing the
value of a like as the revenue coming from consumers who like the firm will overstate
the value of a social media strategy as it ignores all the other costs of servicing these customers. Secondly, it is important to understand that this metric is not the amount that
the organization can spend up to in order to secure a like (follower). Those who interact
(positively) with a firm on social media are likely to already be positively predisposed
to the firm. Encouraging other members of the public, who don’t have the same prior
positive predisposition, by, for example, rewarding those who follow a brand on social
media, is unlikely to change the new recruits’ purchase behaviors to resemble that of the
more avid followers who followed without the incentive. When the difference in value
is not caused by the social media strategy it would be wrong to use this as a benchmark
for social media spending. As such although the Value of a Like is potentially an interesting number, the practical benefits of the metric are more limited than they might
first appear.
354
MARKETING METRICS
10.8 Downloads
Monitoring downloads is a way of tracking engagement with the organization.
Downloads (#) = Number of times that an application or file is downloaded (#)
Downloads reflect the success of organizations at getting their applications distributed
to users.
Purpose: To determine effectiveness in getting applications out to users.
Downloads are a common way for marketers to gain a presence with consumers. This
includes applications for mobile phones, for MP3-style devices, and computers.
Apps for iPhones, software trials, spreadsheets, ring tones, white papers, pictures, and
widgets are examples of downloads. These downloads typically provide a benefit to the
consumer in return for a presence on the device of the user. For instance a weather app
might be branded with the Web site of the weather channel and provide updates on
atmospheric conditions. A consumer packaged goods company might supply an app
that suggests recipes that could use its products in novel ways.
Construction
This metric will be supplied by analytics software.
Downloads (#) = Number of times that an application or file is downloaded (#)
Data Sources, Complications, and Cautions
Downloads is a simple count of the number of times an application or file is downloaded, regardless of who requested the download. It does not distinguish 10 identical
downloads to a given individual from 10 separate downloads to 10 separate individuals,
although these two situations may have dramatically different consequences. In this
way downloads is akin to impressions where a given number of impressions can be
obtained by a variety of combinations of reach and frequency (see Section 9.3). Using
your analytics package to see the number of downloads in unique sessions versus the
total downloads should help you get a better feel for how many downloads were made
by the same person (technically this is often measured as downloads made by the same
browser).
A consideration in the counting of downloads is how to handle downloads that are
started but not completed. One alternative is to keep track of both downloads started
Chapter 10 Online, Email, and Mobile Metrics
355
and downloads completed, another alternative is to pick one or the other (starts or
completions) and use that measure consistently. As always, it is imperative for the user
to know which convention was used in construction of the metric.
A further challenge is that “Downloads” often raise problems in standard Web analytics
packages. This is because they often don’t fire a pageview when they provide a file (e.g.,
a Pdf, doc, xls). This means downloads can’t be tracked like normal “pages.” Downloads
instead need to be tracked with event tracking. Event tracking involves code that creates
a virtual pageview when the event is triggered.
10.9 Mobile Metrics
Mobile metrics may be classified into metrics that are akin to Web site metrics and
other metrics that focus on apps. Average revenue per user (ARPU) is a widely used
metric for app marketers.
Total Revenue ($)
Average Revenue Per User ($) =
Number of Users (#)
Many issues in mobile marketing are similar to marketing aimed at users on computers. (Mobile relates to marketing using platforms optimized for use on hand-held and
other portable devices.) Mobile site metrics are similar to Web site metrics, and can use
Web analytics packages such as the specialized mobile version of metrics on Google
Analytics. These allow you to better understand the users’ interactions and experience
with the mobile site. While the metrics may be similar to traditional Web site measures,
one would expect the results to differ by channel to reflect, for example, the relative
strengths of desktop and mobile search. Mobile marketing might lead to fewer sales on
the devices but might have a greater role in providing support to the sale, such as downloads of coupons or store maps. (Clearly there may be challenges attributing credit for
sales between the mobile site, Web site, and the bricks and mortar store.)
Your analytics package is likely to be able to tell you where your traffic is coming from,
i.e. 80% desktop and 20% mobile. Given the problems making attributions this is only a
useful, but clearly far from complete, view of the reliance of your marketing on various
platforms (the environment the software runs on).
The second strand of mobile marketing might be the use of apps. Ensuring apps are
downloaded is an element of mobile marketing. Once you have an installed user base
you will want to check how you are acquiring customers, how they use the app, and
how long customers stay with you.
356
MARKETING METRICS
With the rise of mobile payments, a marketer will want to be aware of the percentage of
customers using this method of payment. Furthermore marketers will want to monitor
mobile coupons used. The principles are similar to the non-mobile world but the data is
rich, and relatively clean from mobile devices. For example, people usually operate only
their own personal phone whereas people are more likely to share computers.
Construction
We highlight a selection of metrics related to mobile apps.
Session Length: This is the length of time the user spends on the app.
Average Session Length (#) =
Total Length of Time on App (#)
Number of Sessions (#)
Active Users (Monthly/Daily): This is the count of users who use the app during a
given period; for example, a month or a day.
New user acquisition and retention rate are similar to the metrics highlighted in the
customer section of the book (Chapter 5).
Average Revenue Per User: This shows the effectiveness of the marketer at gaining revenue from each user. Clearly the importance of this metric depends upon the purpose
of the app. An app that is designed to build the brand or facilitate usage of a product
may have no revenue associated with it but this does not mean it is not a valuable part
of the marketing strategy.
Average Revenue Per User ($) =
Total Revenue ($)
Number of Users (#)
Store Visits: There are also a number of emerging applications that meld the online
and offline worlds. For example, store visits can be estimated using location-based
tracking on mobile phones. This is anonymized data so you can estimate the number of
consumers who went to the stores—not which customers went there. When you have
the number of consumers going into a store you can compare this to an online action,
for instance the number who downloaded a coupon, to assess how effectively the online
strategy drives offline actions, visits to stores.
Online to Offline Conversion (%) =
Estimated Store Visits (#)
Online Actions (#)
Chapter 10 Online, Email, and Mobile Metrics
357
Data Sources, Complications, and Cautions
Many of the key bits of data needed to assess these metrics come from analytics packages, such as Google’s Mobile analytics package.
While many of these concepts are the same across platforms it is worth noting just how
quickly the mobile market changes. In many areas you might be expecting to see a payback to an investment in a few years. In mobile marketing often payback needs to be
much quicker. If a downloaded app doesn’t pay back in weeks it may never do so. The
active lives of users of mobile apps may be quite short meaning churn will be extremely
high. Indeed you probably want to measure on a much shorter period to make any
retention related metrics more meaningful.
10.10 Email Metrics
Monitoring email response is an important way of judging the success of email campaigns, individual emails, and lists.
Email metrics tend to measure the response of consumers to the email. Does the email
lead to an action? For example, is an email opened?
Email Open Rate (%) =
Emails Opened (#)
Emails Delivered (#)
Bounce rate in email marketing is a measure of the quality of the list and is essentially independent of consumer action. This is a notable difference from bounce rate
when measuring a Web site, which measures the action the Web site generates from
consumers.
Purpose: To determine effectiveness of email campaigns
Despite the popularity of search, email marketing retains a place in the marketing
endeavors of many firms. Many of these metrics appear similar to Web site metrics as
they have a similar purpose. The metrics are focused on measuring the level of activity
that is generated by the marketer’s actions.
The simplest metric is email open rate, which is the percentage of emails that get
opened. Clearly emails that don’t get opened are relatively unlikely to motivate any
further action.
358
MARKETING METRICS
Clickthrough Rate is similar to the search metrics and represents the percentage of
emails that are clicked on. These might typically bring the consumer through to a Web
site where they could buy a product.
Email can annoy people and unsubscribes captures the percentage of subscribers who
ask to be removed from the list. All else equal a list with a high unsubscribe rate is a
worse list. (Although a list on which consumers continue to receive email but aren’t
motivated to any action is not especially useful. It is simply a wasted expense to create
the emails.)
Construction
Email Open Rate: The percentage of email delivered that gets opened.
Email Open Rate (%) =
Emails Opened (#)
Emails Delivered (#)
Email Clickthrough Rate: The percentage of email delivered that gets clicked on.
Email Clickthrough Rate (%) =
Emails Clicked (#)
Emails Delivered (#)
Email Unsubscribe Rate: The percentage of any list of email subscribers that opt out of
the list in any given period.
Email Unsubscribe Rate (%) =
Requests to Be Unsubscribed (#)
Total Subscribers at Beginning of Period (#)
Bounce Rate: One notable complication is that email bounce rate has a different definition than Web site bounce rate. Email bounce rate records the quality of your list rather
than any action on the part of the recipient in respect of the email. An older, poorer
quality list is likely to have a higher bounce rate as a percentage of recipients will have
abandoned their old email addresses. (Email bounce rate relates to emails that can’t be
read.) Investigating the data further can let the marketer know more about the source of
failure, e.g., did the recipient’s server accept the email before bouncing it back, perhaps
because the email address has reached its maximum size?
Email Bounce Rate (%) =
Number of Emails Not Delivered (#)
Total Emails Sent (#)
Cost Per Engagement: Marketers often refer to “cost per engagement” (CPE). This is
a similar metric to the other cost per metrics we have discussed. The only additional
consideration is that we must define exactly what an engagement is. An engagement
Chapter 10 Online, Email, and Mobile Metrics
359
is essentially whatever the marketer decides based upon the goals of the campaign. A
marketer trying to gain email newsletter signups might pay a third party on a Cost Per
Engagement basis, paying only for signups. Clearly the third parties must be willing to
work closely with marketers to establish the definition of engagement, and associated
tracking mechanisms.
Cost per Engagement ($) =
Cost of Activity ($)
Number of Engagements (#)
Data Sources, Complications, and Cautions
The rules on what information can be held by marketers on consumers vary by jurisdiction. Many parts of the world require consent of the recipient to receiving email.
You should always check that you are complying with local laws and best practice.
The Direct Marketing Association (http://thedma.org) and American Marketing
Association (https://www.ama.org) have helpful resources available.
There can be a considerable difference between emails sent and delivered given abandoned email addresses, bounced emails, and full inboxes. Even those that are delivered
may be filtered away from the recipient’s attention. Email marketers should seek to
understand the fate of emails they send in greater detail. Increasing spam and other
filtering on the part of the users can lead to reduced rates that have nothing to do with
the effectiveness of the creative message sent. The recipients may simply never see the
message so it cannot have an effect even if the message is powerful.
The effect of a campaign can be monitored by tagging URLs in emails. Using the URL
one can see whether the message drives traffic to the site and which sites have traffic
that results in conversions.
Email opens and clicks can be used in testing different emails. When testing, e.g., an
A/B test, a marketer sends out two types of email. In general for testing one should limit
the differences between the emails. This difference constitutes the thing being tested
and so if you vary many things it requires more advanced stats to tease out the effects.
(And sometimes teasing out the effects is simply impossible where, for example, two
changes always occur together it is impossible to assess the effect of each change independently.) For example, one email creative may offer: “Buy One, Get One Free” while
the other creative offers: “Two For The Price of One.) Sending these emails randomly
to the email list allows the marketer to assume that the recipients of each email were
similar. Armed with this assumption the marketer can then determine which creative,
“two for the price of one” or “buy one get one free,” is more effective based upon which
is more likely to be opened and clicked on.
360
MARKETING METRICS
Email marketing relies on the quality of the email list. The quality of lists, for a specific
task, can be tested by observing the open and click rate of a message sent to several lists.
High quality lists can still gain little response if the offer sent is inappropriate to the
members of the list.
References and Suggested Further Reading
Google Analytics Academy, https://analyticsacademy.withgoogle.com/explorer. Accessed
3/2/2015.
Kaushik, Avinash (2007) Web Analytics: An Hour A Day, Sybex.
Net Genesis Corp. (2000). E-metrics Business Metrics for the New Economy. Net Genesis and
Target Marketing of Santa Barbara.
Pirouz, Raymond, Neil Bendle, and Taneem Khan (2011) Online Metrics: What Are You
Measuring and Why? Ivey Business School, 9B11A016.
Chapter 10 Online, Email, and Mobile Metrics
361
This page intentionally left blank
11
MARKETING AND FINANCE
Introduction
Key concepts covered in this chapter:
Net Profit and Return on Sales (ROS)
Project Metrics: Payback, NPV, IRR
Return on Investment (ROI)
Marketing ROI (MROI or ROMI)
Economic Profit (aka, EVA®)
As marketers progress in their careers, it becomes increasingly necessary to coordinate their plans with other functional areas. Sales forecasts, budgeting, and estimating
returns from proposed marketing initiatives are often the focus of discussions between
marketing and finance. For marketers with little exposure to basic finance metrics, a
good starting point is to gain a deeper understanding of “rate of return.” “Return” is
generally associated with profit, or at least positive cash flow. “Return” also implies
that something has left—cash outflow. Almost all business activity requires some cash
outflow. Even sales cost money that is only returned when bills are paid. In this chapter we provide a brief overview of some of the more commonly employed measures of
profitability and profits. Understanding how the metrics are constructed and used by
finance to rank various projects will make it easier to develop marketing plans that meet
the appropriate criteria.
The first section covers net profits and return on sales (ROS). Next, we look at return
on investment (ROI), the ratio of net profit to amount of investment. Another metric
that accounts for the capital investment required to earn profits is economic profits
(also known as economic value added—EVA), or residual income. Because EVA and
ROI provide snapshots of the per-period profitability of firms, they are not appropriate for valuing projects spanning multiple periods. For multi-period projects, three of
363
the most common metrics are payback, net present value (NPV), and internal rate of
return (IRR).
The last section discusses the frequently mentioned but rarely defined measure,
Marketing Return on Investment. Although this is a well-intentioned effort to measure
marketing productivity, consensus definitions and measurement procedures for “marketing ROI” or MROI (aka ROMI) have yet to emerge.
Metric
Construction
11.1
Net Profit
Sales revenue less Revenue and costs The basic profit
total costs.
can be defined
equation.
in a number of
ways leading to
confusion in profit
calculations.
11.1
Return on Sales
(ROS)
Net profit as a
percentage of
sales revenue.
Acceptable level
of return varies
between industries
and business
models. Many
models can be
described as high
volume/low return
or vice versa.
11.1
Earnings Before
Interest, Taxes,
Depreciation,
and Amortization
(EBITDA)
Earnings Before
Interest, Taxes,
Depreciation, and
Authorization.
Strips out the effect Rough measure
of accounting and of operating cash
financing polices
flow.
from profits.
Ignores important
factors, such as
depreciation of
assets.
11.2
Return on
Investment (ROI)
Net profits over
the investment
needed to
generate the
profits.
Often meaningless
in the short term.
Variations such as
return on assets
and return on
investment capital
analyze profits
in respect of
different inputs.
364
MARKETING METRICS
Considerations
Purpose
Gives the
percentage of
revenue that is
being captured in
profits.
A metric that
describes how well
assets are being
used.
Metric
Construction
Considerations
Purpose
11.3
Economic Profit
(aka EVA®,
Economic Value
Added)
Net operating
profit after tax
(NOPAT) less the
cost of capital.
Requires a cost
of capital to
be provided/
calculated.
Shows profit
made in dollar
terms. Gives a
clearer distinction
between the sizes
of returns than
does a percentage
calculation.
11.4
Payback
The length of
time taken to
return the initial
investment.
Will favor projects
with quick returns
more than longterm success.
Simple return
calculation.
11.4
Net Present Value
(NPV)
The value of a
stream of future
cash flows after
accounting for
the time value of
money.
The discount rate
used is the vital
consideration
and should account
for the risk of the
investment
too.
To summarize
the value of cash
flows over multiple
periods.
11.4
Internal Rate of
Return (IRR)
The discount rate
at which the NPV
of an investment
is zero.
IRR does not
describe the
magnitude of
return; $1 on $10
is the same as $1
million on $10
million.
An IRR will
typically be
compared to a
firm’s hurdle rate.
If IRR is higher
than hurdle rate,
invest; if lower,
pass.
11.5
Marketing Return
on Investment;
Revenue (MROI
or ROMI)
Incremental
financial impact
attributable
to marketing
divided by
the marketing
spending.
Marketers need
to establish an
accurate baseline
to be able to
meaningfully
estimate
incremental
financial effects
of marketing.
Evaluates
productivity of
various marketing
efforts. The
percentage term
helps comparison
across plans
of varying
magnitude.
Chapter 11 Marketing and Finance
365
11.1 Net Profit and Return on Sales
Net profit measures the profitability of ventures after accounting for all costs. Return
on sales (ROS) is net profit as a percentage of sales revenue.
Net Profit ($) = Sales Revenue ($) – Total Costs ($)
Return on Sales—ROS (%) =
Net Profit ($)
Sales Revenue ($)
EBITDA ($) = Net Profit ($) + Interest Payments ($) + Taxes ($) +
Depreciation and Authorization Charges ($)
ROS is an indicator of profitability and is often used to compare the profitability of
companies and industries of differing sizes. Significantly, ROS does not account for
the capital (investment) used to generate the profit.
Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) is a rough
measure of operating cash flow, which reduces the effect of accounting, financing, and
tax polices on reported profits.
Purpose: To measure levels and rates of profitability.
How does a company decide whether it is successful or not? Probably the most common
way is to look at the net profits of the business. Given that companies are collections of
projects and markets, individual areas can be judged on how successful they are at adding to the corporate net profit. Not all projects are of equal size, however, and one way
to adjust for size is to divide the profit by sales revenue. The resulting ratio is return on
sales (ROS), the percentage of sales revenue that gets “returned” to the company as net
profits after all the related costs of the activity are deducted.
Construction
Net profit measures the fundamental profitability of the business. It is the revenues of
the activity less the costs of the activity. The main complication is in more complex
businesses when overhead needs to be allocated across divisions of the company (see
Figure 11.1). Almost by definition, overheads are costs that cannot be directly tied to any
specific product or division. The classic example would be the cost of headquarters staff.
Net Profit: To calculate net profit for a unit (such as a company or division), subtract all costs, including a fair share of total corporate overheads, from the gross
revenues.
Net Profit ($) = Sales Revenue ($) – Total Costs ($)
366
MARKETING METRICS
Sales Revenues for the Firm
Total Variable Costs
Line Specific
Fixed Costs
Overhead
Business
Net Profit
Simple View of Business – Revenues and Costs
Figure 11.1
Profits = Revenues Less Costs
Return on Sales (ROS): Net profit as a percentage of sales revenue.
Return on Sales (%) =
Net Profit ($)
Sales Revenue ($)
Earnings before interest, taxes, depreciation, and amortization (EBITDA) is a very
popular measure of financial performance. It is used to assess the “operating” profit of
the business. It is a rough way of calculating how much cash the business is generating and is even sometimes called the “operating cash flow.” It can be useful because it
removes factors that change the view of performance depending upon the accounting
and financing policies of the business. Supporters argue it reduces management’s ability
to change the profits they report by their choice of accounting rules and the way they
generate financial backing for the company. This metric excludes from consideration
expenses related to decisions such as how to finance the business (debt or equity) and
over what period to depreciate fixed assets. EBITDA is typically closer to actual cash
flow than is NOPAT (discussed later in the chapter).
EBITDA can be calculated by adding back the costs of interest, depreciation, and amortization charges and any taxes incurred.
EBITDA ($) = Net Profit ($) + Interest Payments ($) + Taxes Incurred ($) + Depreciation
and Amortization Charges ($)
Data Sources, Complications, and Cautions
Although it is theoretically possible to calculate profits for any sub-unit, such as a product
or region, often the calculations are rendered suspect by the need to allocate overhead
costs. Because overhead costs often don’t come in neat packages, their allocation among
the divisions or product lines of the company can often be more art than science.
Chapter 11 Marketing and Finance
367
For return on sales, it is worth bearing in mind that a “healthy” figure depends on the
industry and capital intensity (amount of assets per sales dollar). Return on sales is similar
to margin (%), except that ROS accounts for overheads and other fixed costs that are often
ignored when calculating margin (%) or contribution margin (%). (Refer to Section 3.1.)
Related Metrics and Concepts
Net operating profit after tax (NOPAT) deducts relevant income taxes but excludes
some items that are deemed to be unrelated to the main (“operating”) business.
11.2 Return on Investment
Return on investment is one way of considering profits in relation to capital
invested.
Net Profit ($)
Return on Investment—ROI (%) =
Investment ($)
Return on assets (ROA), return on net assets (RONA), return on capital (ROC),
and return on invested capital (ROIC) are similar measures with variations on how
“investment” is defined.
Marketing not only influences net profits but also can affect investment levels too.
New plants and equipment, inventories, and accounts receivable are three of the
main categories of investments that can be affected by marketing decisions.
Purpose: To measure per period rates of return on dollars invested
in an economic entity.
ROI and related metrics (ROA, ROC, RONA, and ROIC) provide a snapshot of profitability adjusted for the size of the investment assets tied up in the enterprise. Marketing
decisions have obvious potential connection to the numerator of ROI (profits), but
these same decisions often influence assets usage and capital requirements (for example, receivables and inventories). Marketers should understand the position of their
company and the returns expected. ROI is often compared to expected (or required)
rates of return on dollars invested.
Construction
For a single period review just divide the return (net profit) by the resources that were
committed (investment):
Return on Investment (%) =
368
MARKETING METRICS
Net Profit ($)
Investment ($)
Data Sources, Complications, and Cautions
Averaging the profits and investments over periods such as one year can disguise wide
swings in profits and assets, especially inventories and receivables. This is especially
true for seasonal businesses (such as some construction materials and toys). In such
businesses it is important to understand these seasonal variations to relate quarterly
and annual figures to each other.
Related Metrics and Concepts
Return on assets (ROA), return on net assets (RONA), return on capital employed
(ROCE), and return on invested capital (ROIC) are commonly used variants of ROI.
They are also calculated using net profit as the numerator, but they have different
denominators. The relatively subtle distinctions between these metrics are beyond
the scope of this book. Some differences are found in whether payables are subtracted
from working capital and how borrowed funds and stockholder equity are treated.
11.3 Economic Profit—EVA
Economic profit has many names, some of them trademarked as “brands.”
Economic value added (EVA) is Stern-Stewart’s trademark. They deserve credit for
popularizing this measure of net operating profit after tax adjusted for the cost of
capital.
Economic Profit ($) = Net Operating Profit After Tax (NOPAT) ($) – Cost of Capital ($)
Cost of Capital ($) = Capital Employed ($) * WACC (%)
Unlike percentage measures of return (for example, ROS or ROI), Economic profit
is a dollar metric. As such, it reflects not only the “rate” of profitability, but also the
size of the business (sales and assets).
Purpose: To measure dollar profits while accounting
for required returns on capital invested.
Economic profit, sometimes called residual income, or EVA, is different from
“accounting” profit, in that economic profit also considers the cost of invested capital—
the opportunity cost (see Figure 11.2). Like the discount rate for NPV calculations, this
charge should also account for the risk associated with the investment. A popular (and
proprietary) way of looking at economic profit is economic value added.1
Increasingly, marketers are being made aware of how some of their decisions influence
the amount of capital invested or assets employed. First, sales growth almost always
requires additional investment in fixed assets, receivables, or inventories. Economic
Chapter 11 Marketing and Finance
369
profit and EVA help determine whether these investments are justified by the profit
earned. Second, the marketing improvements in supply chain management and channel coordination often show up in reduced investments in inventories and receivables.
In some cases, even if sales and profit fall, the investment reduction can be worthwhile. Economic profit is a metric that will help assess whether these trade-offs are
being made correctly.
After-Tax
Operating Profit
EVA
Minus
A Charge for
Capital Used
Figure 11.2
EVA Is After-Tax Profit Minus a Charge for Capital Usage
Construction
Economic profit/economic value added can be calculated in three stages. First, determine NOPAT (net operating profit after tax). Second, calculate the cost of capital by
multiplying capital employed by the weighted average cost of capital.2 The third stage is
to subtract the cost of capital from NOPAT.
Economic Profit ($) = Net Operating Profit After Tax (NOPAT) ($) – Cost of Capital ($)
Cost of Capital ($) = Capital Employed ($) * WACC (%)
Economic Profit: If your profits are less than the cost of capital, you have
lost value for the firm. Where economic profit is positive, value has been
generated.
EXAMPLE: A company has profits—NOPAT—of $145,000.
They have a straightforward capital structure, half of which is supplied by shareholders. This equity
expects a 12% return on the risk the shareholders are taking by investing in this company. The other
half of the capital comes from a bank at a charge of 6%:
370
MARKETING METRICS
Weighted average cost of capital (WACC) is therefore
= Equity (12% * 50%) + Debt (6% * 50%) = 9%
The company employs total capital of $1 million. Multiplying the capital employed by the weighted
average cost for the capital employed will give us an estimate of the profit (return) required to cover
the opportunity cost of capital used in the business:
Cost of Capital = Capital Employed * WACC
= $1,000,000 * 9%
= $90,000
Economic profit is the surplus of profits over the expected return to capital.
Economic Profit = NOPAT – Cost of Capital
= $145,000 – $90,000
= $55,000
Data Sources, Complications, and Cautions
Economic profit can give a different ranking for companies than does return on investment. This is especially true for companies such as Wal-Mart and Microsoft that have
experienced (achieved) high rates of growth in sales. Judging the results of the giant
U.S. retailer Wal-Mart by many conventional metrics will disguise its success. Although
the rates of return are generally good, they hardly imply the rise to dominance that the
company achieved. Economic profit reflects both Wal-Mart’s rapid sales growth and
its adequate return on the capital invested. This metric shows the magnitude of profits after the cost of capital has been subtracted. This combines the idea of a return on
investment with a sense of volume of profits. Simply put, Wal-Mart achieved the trick
of continuing to gain decent returns on a dramatically increasing pool of capital.
11.4 Evaluating Multi-period Investments
Multi-period investments are commonly evaluated with three metrics.
Payback (#) = The number of periods required to “pay back”
or “return” the initial investment.
Net Present Value (NPV) ($) = The discounted value of future cash flows
minus the initial investment.
Internal Rate of Return (IRR) (%) = The discount rate that results in an NPV of zero.
Chapter 11 Marketing and Finance
371
These three metrics all deal with economic consequences occurring at different
points in time.
Purpose: To evaluate investments with financial consequences
spanning multiple periods.
Investment is a word business people like. It has all sorts of positive connotations of
future success and wise stewardship. However, because not all investments can be pursued, those available must be ranked against each other. Also, some investments are not
attractive even if we have enough cash to fund them. In a single period, the return on any
investment is merely the net profits produced in the time considered divided by the capital invested. Evaluation of investments that produce returns over multiple periods requires
a more complicated analysis—one that considers both the magnitude and timing of the
returns.
Payback (#): The time (usually years) required to generate the (undiscounted)
cash flow to recover the initial investment.
Net Present Value—NPV ($): The present (discounted) value of future cash
inflows minus the present value of the investment and any associated future cash
outflows.
Internal Rate of Return—IRR (%): The discount rate that results in a net present value of zero for a series of future cash flows after accounting for the initial
investment.
Construction
Payback: The years required for an investment to return the initial investment.
Projects with a shorter payback period by this analysis are regarded more favorably
because they allow the resources to be reused quickly. Also, generally speaking, the
shorter the payback period, the less uncertainty is involved in receiving the returns. Of
course the main flaw with payback period analysis is that it ignores all cash flows after
the payback period. As a consequence, projects that are attractive but that do not produce immediate returns will be penalized with this metric.
EXAMPLE: Harry is considering buying a small chain of hairdressing salons. He estimates
that the salons will produce a net income of $15,000 a year for at least five years. Harry’s payback on this investment is $50,000/$15,000, or 3.33 years.
372
MARKETING METRICS
Net Present Value
Net present value (NPV) is the discounted value of the cash flows associated with the
project.
The present value of a dollar received in a given number of periods in the future is
Cash Flow ($) * 1
Discounted Value ($) =
(1 + Discount Rate (%)) ^ Period (#)
This is easiest to see when set out in spreadsheet form.
A 10% discount rate applied to $1 received now and in each of the next three years
reduces in value over time as shown in Table 11.1.
Table 11.1
Discounting Nominal Values
Year 0
Year 1
Year 2
Year 3
Discount Formula
1
1/(1+10%)^1
1/(1+10%)^2
1/(1+10%)^3
Discount Factor
1
90.9%
82.6%
75.1%
Undiscounted Cash
Flows
$1.00
$1.00
$1.00
$1.00
Present Value
$1.00
$0.91
$0.83
$0.75
Spreadsheets make it easy to calculate the appropriate discount factors.
EXAMPLE: Harry wants to know the dollar value of his business opportunity. Although he is
confident about the success of the venture, all future cash flows have a level of uncertainty. After
receiving a friend’s advice, he decides a 10% discount rate on future cash flows is about right.
He enters all the cash flow details into a spreadsheet (see Table 11.2).3 Harry works out the discount
factor using the formula and his discount rate of 10%:
Discounted Value =
For Year 1 Cashflows =
Cash Flow * 1
(1 + Discount Rate) ^ Year
$15,000 * 1
$15,000 * 1
=
(1 + 10%) ^ 1)
110%
= $15,000 * 90.9% = 13,636
Chapter 11 Marketing and Finance
373
Table 11.2
Year 0
Investment
Discounted Cashflow (10% Discount Rate)
Year 1
Year 2
Year 3
Year 4
Year 5
($50,000)
Total
($50,000)
Income
$15,000
$15,000 $15,000
$15,000
$15,000
$75,000
Undiscounted ($50,000)
Cashflow
$15,000
$15,000 $15,000
$15,000
$15,000
$25,000
Discount
Formula
1/(1+
DR)^0
1/(1+
DR)^1
1/(1+
DR)^2
1/(1+
DR)^3
1/(1+
DR)^4
1/(1+
DR)^5
Discount
Factor
100.0%
90.9%
82.6%
75.1%
68.3%
62.1%
$13,636
$12,397 $11,270
$10,245
$9,314
Present Value ($50,000)
$6,862
The NPV of Harry’s project is $6,862. Of course the NPV is lower than the sum of the undiscounted
cash flows. NPV accounts for the fact that on a per-dollar basis, cash flows received in the future are
less valuable than cash in hand.
Internal Rate of Return
The internal rate of return is the percentage return made on the investment over a
period of time. The internal rate of return is a feature supplied on most spreadsheets
and thus is relatively easy to calculate.
Internal Rate of Return (IRR): The discount rate for which the net present
value of the investment is zero.
The IRR is especially useful because it can be compared to a company’s hurdle rate.
The hurdle rate is the necessary percentage return to justify a project. Thus a company
might decide only to undertake projects with a return greater than 12%. Projects that
have an IRR greater than 12% get the green light; all others are thrown in the bin.
EXAMPLE: Returning to Harry, we can see that IRR is an easy calculation to perform using a
software package. Enter the values given in the relevant periods on the spreadsheet (see Table 11.3).
Year 0—now—is when Harry makes the initial investment; each of the next five years sees a $15,000
return. Applying the IRR function gives a return of 15.24%.
374
MARKETING METRICS
Table 11.3 Five-Year Cashflow
Cell ref
A
1
2
Cashflows
B
C
D
E
F
G
Year 0
Year 1
Year 2
Year 3
Year 4
Year 5
($50,000)
$15,000
$15,000
$15,000
$15,000
$15,000
In Microsoft Excel, the function is = IRR(B2:G2)
which equals 15.24%.
The cell references in Table 11.3 should help in re-creating this function. The function is telling
Excel to perform an IRR on the range B2 (cashflow for year 0) to G2 (cashflow for year 5).
IRR and NPV Are Related
The internal rate of return is the percentage discount rate at which the net present value
of the operation is zero.
Thus companies using a hurdle rate are really saying that they will only accept projects
where the net present value is positive at the discount rate they specify as the hurdle
rate. Another way to say this is that they will accept projects only if the IRR is greater
than the hurdle rate.
Data Sources, Complications, and Cautions
Payback and IRR calculations require estimates of cash flows. The cash flows are the
monies received and paid out that are associated with the project per period, including
the initial investment. Topics that are beyond the scope of this book include the time
frame over which forecasts of cash flows are made and how to handle “terminal values”
(the value associated with the opportunity at the end of the last period).4 Net present
value calculations require the same inputs as payback and IRR, plus one other: the
discount rate. Typically, the discount rate is decided at the corporate level. This rate has
a dual purpose to compensate for the following:
■
The time value of money
■
The risk inherent in the activity
A general principle to employ is that the riskier the project, the greater the discount rate
to use. Considerations for setting the discounts rates are also beyond the scope of this
book. We will simply observe that, ideally, separate discount rates would be assessed
Chapter 11 Marketing and Finance
375
for each individual project because risk varies by activity. A government contract might
be a fairly certain project—not so for an investment by the same company in buying a
fashion retailer. The same concern occurs when companies set a single hurdle rate for
all projects assessed by IRR analysis.
Cashflows and Net Profits: In our examples cash flow equals profit, but in
many cases they will be different.
A Note for Users of Spreadsheet Programs
Microsoft Excel has an NPV calculator, which can be very useful in calculating NPV.
The formula to use is NPV(rate,value1,value2, etc.) where the rate is the discount
rate and the values are the cash flows by year, so year 1 = value 1, year 2 = value 2,
and so on.
The calculation starts in period one, and the cash flow for that period is discounted. If you are using the convention of having the investment in the period
before, i.e. period 0, you should not discount it but add it back outside the formula.
Therefore Harry’s returns discounted at 10% would be
= NPV(Rate, Value 1, Value 2, Value 3, Value 4, Value 5)
= NPV(10%, 15000, 15000, 15000, 15000, 15000) or $56,861.80 less the
initial investment of $50,000.
This gives the NPV of $6,861.80 as demonstrated fully in the example.
11.5 Marketing Return on Investment
Marketing Return on Investment (MROI), also known as (ROMI), is a relatively new
metric. It is not like the other “return-on-investment” metrics because marketing is
not the same kind of investment. Instead of moneys that are “tied” up in plants and
inventories, marketing funds are typically “risked.” Marketing spending is typically
expensed in the current period. There are many variations in the way this metric has
been used, and although no authoritative sources for defining it exist, we believe the
consensus of usage refers to MROI as the dollar-denominated estimate of the incremental financial value to the entity generated by identifiable marketing expenditures,
less the cost of those expenditures as a percentage of the same expenditures:
Marketing Return on
=
Investment (MROI) (%)
376
MARKETING METRICS
Incremental Financial Value Created By
Marketing ($) – Cost of Marketing ($)
Cost of Marketing ($)
There are many ways to estimate the financial value generated by marketing. The
most commonly employed method is to estimate incremental contribution margin,
net of marketing, generated by marketing divided by the marketing spend.
Purpose: To measure the rate at which spending on marketing contributes
to profits.
Marketers are under more and more pressure to “show a return” on their activities.
However, it is often unclear exactly what this means. Certainly, marketing spending is
not an “investment” in many senses of the word. There is usually no tangible asset and
often not even a predictable (quantifiable) result to show for the spending, but marketers still want to emphasize that their activities contribute to financial health. Some
might argue that marketing should be considered an expense and the focus should be
on whether it is a necessary expense. Marketers, however, typically believe that many
of their activities generate lasting results and therefore should be considered “investments” in the future of the business.5
Marketing Return on Investment (MROI), also known as Return on
Marketing Investment (ROMI): The incremental financial value attributable
to marketing (net of marketing spending), divided by the marketing “invested”
or risked.
Construction
A necessary step in calculating MROI is the estimation of the incremental financial
value attributable to marketing. This incremental value can be “total” attributable to
marketing or the marginal value attributable to a new marketing initiative. The following example, in Figure 11.3, should help clarify the difference:
Y0 = Baseline financial value (with $0 marketing spending),
Y1 = Financial value at marketing spending level X1, and
Y2 = Financial value at marketing spending level X2,
where the difference between X1 and X2 represents the cost of an incremental marketing
budget item that is to be evaluated, such as an advertising campaign or a trade show.
1. Financial Value Attributable to Marketing = Y2 – Y0: The increase in financial
value attributable to the entire marketing budget (equal to financial value with
marketing minus baseline financial value).
Chapter 11 Marketing and Finance
377
2. Marketing Return on Investment (MROI) = [(Y2 – Y0) – X2]/X2: The financial
value created by all marketing activities divided by the cost of those activities.
3. Return on Incremental Marketing Investment (ROIMI) = [(Y2 – Y1) –
(X2 – X1)]/(X2 – X1): The incremental financial value due to the incremental
marketing spending divided by the amount of incremental spending.
Y2
Financial Value ($)
Y1
Y0
X0
X1
X2
Marketing Spending ($)
Figure 11.3
Evaluating the Return to Marketing
EXAMPLE: A farm equipment company was considering a direct mail campaign to
remind customers to have tractors serviced before spring planting. The campaign is expected to cost $1,000 and to increase revenues from $45,000 to $50,000. Baseline revenues for
tractor servicing (with no marketing) were estimated at $25,000. The direct mail campaign was
in addition to the regular advertising and other marketing activities costing $6,000. The firm uses
contribution to assess financial value and the contribution on tractor servicing revenues (after
parts and labor) averages 60%.
For some industries revenue-based metrics might be useful, but for most situations
these metrics are liable to be very misleading. MROI or ROIMI (see following examples) are generally more useful. There is no point in spending $20,000 on advertising
to generate $100,000 of sales—a respectable 500% return to revenue—if high variable
378
MARKETING METRICS
costs mean the marketing only generates a contribution of $5,000. We often calculate
financial value from contribution. Later we discuss other potential approaches.
Marketing Return on Investment =
– MROI (%)
Incremental Financial Value Created By Marketing
($) – Cost of Marketing ($)
Cost of Marketing ($)
EXAMPLE: Each of the metrics in this section can be calculated from the information in the
example. Let us calculate the financial value with the marketing and direct mail (Y2 = Revenue
($50,000) times contribution margin (60%) =$30,000.
The financial value with all the marketing except the direct mail (Y1 = Revenue ($45,000) times
contribution margin (60%) = $27,000).
The financial value with no marketing (Y0 = Revenue ($25,000) times contribution margin (60%)
= $15,000.
The financial value attributable to marketing (Y2 – Y0) = $30,000 – $15,000 = $15,000.
The financial value attributable to the direct mail (Y2 – Y1) = $30,000 – $27,000 = $3,000.
Marketing Return on Investment (MROI) = [Incremental Financial Value Attributable to
Marketing ($) Less Cost of Marketing ($)]/ Cost of
Marketing ($) = ($15,000-$7,000)/$7,000 = 114%
If the direct mail campaign were not used this would be ($12,000-$6,000)/$6,000 = 100%
Return on Incremental
Marketing Investment =
(ROIMI), the direct mail
[Incremental Financial Value Attributable to the
Direct Mail ($) Less Cost of Direct Mail ($)]
($3,000 – $1,000)/$1,000=
= 200%
Cost of Direct Mail ($)
Data Sources, Complications, and Cautions
The first piece of information needed for marketing ROI is the cost of the marketing
campaign, program, or budget. Although defining which costs belong in marketing can
be problematic, a bigger challenge is estimating the incremental revenue, contribution,
and net profits attributable to marketing. This is similar to the distinction between
baseline and lift discussed in Section 8.1. Farris and his colleagues (2015) outline five
ways that financial returns are often assessed; see Table 11.4.
Chapter 11 Marketing and Finance
379
Table 11.4
Approaches to Measuring Financial Return
Valuation Methods
Financial Return Assessed
Comparable Costs
Cost savings for achieving equivalently valuable contacts
Funnel Conversions
Future period incremental sales and profits based on
estimated conversion rates
Baseline-Lift
Current period incremental sales and profits
Customer Equity
Changes in customer lifetime value
Marketing Assets
Changes in brand and firm valuations
We suggest that it is important to understand the approach being used and that it is
appropriate in this situation.
A further complication of estimating MROI concerns how to deal with important
interactions between different marketing programs and campaigns. The return on
many marketing “investments” is likely to show up as an increase in the responses
received for other types of marketing. For example, if direct mail solicitations show
an increase in response because of television advertising, we could and should
calculate that those incremental revenues had something to do with the TV campaign. As an interaction, however, the return on advertising would depend on what
was being spent on other programs. The function is not a simple linear return to the
campaign costs.
For budgeting, one key element to recognize is that maximizing the MROI would probably reduce spending and profits. Marketers typically encounter diminishing returns,
in which each incremental dollar will yield lower and lower incremental MROI, and
so low levels of spending will tend to have very high return rates. Maximizing MROI
might lead to reduced marketing and eliminating campaigns or activities that are,
on balance, profitable, even if the return rates are not as high. This issue is similar
to the distinction between ROI (%) and EVA ($) discussed in Sections 11.2 and 11.3.
Additional marketing activities or campaigns that bring down average percentage
returns but increase overall profits can be quite sensible. So, using MROI or any
percentage measure of profit to determine overall budgets is questionable. Of course,
merely eliminating programs with a negative MROI is almost always a good idea. Of
course, maximizing long-term profits is often not simply a matter of shifting funds
from low ROI to high ROI activities, because there may well be strategic considerations not fully captured in the ROI measures themselves. Examples are brand building and new customer acquisition versus the need for short-term sales, balancing push
and pull efforts to support distribution channels, and targeting market segments that
are of strategic importance.
380
MARKETING METRICS
The previous discussion intentionally does not deal with carryover effect; that is, marketing effects on sales and profits that extend into future periods. When marketing
spending is expected to have effects beyond the current period, other techniques will be
needed. These include payback, net present value, and internal rate of return. Also, see
customer lifetime value (Section 5.3) for a more disaggregated approach to evaluating
marketing spending designed to acquire long-lived customer relationships. All of the
above discussion is based on referring to the “R” in MROI as “return of profit” resulting from marketing expenditures. The reality is that several firms today are now using
return in various different dimensions other than just profit contributions. While this
can be confusing, part of the reason for this might be because the immediate profits
are less obvious, costs are unknown, or the associated incremental revenue is not easy
to determine. Hence, many are referring to “return” per dollar spent as a way to judge
achieving some interim stage in the purchase funnel as described below.
Related Metrics
Media Exposure Return on Marketing Investment: In an attempt to evaluate the value
of marketing activities such as sponsorships, marketers often commission research to
gauge the number and quality of media exposures achieved. These exposures are then
valued (often using “rate cards” to determine the cost of equivalent advertising space/
time) and a “return” is calculated by dividing the estimated value by the costs.
Media Exposure Return on Marketing =
Investment – (MEROMI) (%)
Estimated Value of Media Exposures Achieved ($)
– Cost of Marketing Campaign, Sponsorship,
or Promotion ($)
Cost of Marketing Campaign, Sponsorship,
or Promotion ($)
This is most appropriate where there isn’t a clear market rate for the results of the campaign and so marketers want to be able to illustrate the equivalent cost for the result for
a type of campaign that has an established market rate.
EXAMPLE: A travel portal decides to sponsor a car at a Formula 1 event. It assumes that the
logo it puts on the car will gain the equivalent of 500,000 impressions and will cost 10,000,000 yen.
The cost per impression is thus 10 million yen/500,000 = or 20 yen per impression. This can be
compared to the costs of other marketing campaigns.
Because MROI has been used in so many different contexts, (incremental versus marginal or total spending, short term versus long term, and with the “R” expressed not just
in terms of profit but as some level in the purchase funnel), MROI can be confusing.
Chapter 11 Marketing and Finance
381
Farris, et al. (2015) suggest that for clarification all expressions of MROI should be
expressed as, “Our analysis measured a (total, incremental, or marginal) MROI of
(scope of spending) using (valuation method) over time period.”
References and Suggested Further Reading
Farris, Paul W., Dominique M. Hanssens, James D. Lenskold, and David J. Reibstein, “Marketing
Return on Investment: Seeking Clarity for Concept and Measurement,” Applied Marketing
Analytics, Vol. 1, Number 3, July 2015.
Hawkins, D. I., Roger J. Best, and Charles M. Lillis. (1987). “The Nature and Measurement of
Marketing Productivity in Consumer Durables Industries: A Firm Level Analysis,” Journal of the
Academy of Marketing Science, 1(4), 1–8.
382
MARKETING METRICS
12
THE MARKETING METRICS
X-RAY AND TESTING
12.1 The Marketing Metrics X-Ray
Our purpose in this chapter is to give some examples of how marketing metrics can
augment and complement traditional financial metrics when used to assess firm and
brand performance. In particular, marketing metrics can serve as leading indicators of
problems, opportunities, and future financial performance. Just as x-rays (now MRIs)
are designed to provide deeper views of our bodies, marketing metrics can show problems (and opportunities) that would otherwise be missed.
Put Your Money Where Your Metrics Are
Table 12.1 shows common summary financial information for two hypothetical companies, Boom and Cruise. Income statement data from five years provide the basis for
comparing the companies on several dimensions.
On Which Firm Would You Bet Your Grandparents’ Savings?
We have used this example with MBA students and executives many times—usually,
we ask them “Assume that your grandparent wants to buy a partnership in one of these
firms, using limited retirement savings. If these financial statements were the only data
you had available or could obtain, which firm would you recommend?” These data are
the metrics traditionally used to evaluate firm performance.
The table shows that gross margins and profits are the same for both firms. Although
Boom’s sales and marketing spending are growing faster, its return on sales (ROS) and
return on investment (ROI) are declining. If this decline continues, Boom will be in
trouble. In addition, Boom’s marketing/sales ratio is increasing faster than Cruise’s. Is
this a sign of inefficient marketing?
383
Table 12.1 Financial Statements
Boom
All $ in (Thousands)
Year 1
Year 2
Year 3
Year 4
Revenue
$833
$1,167
$1,700
$2,553
$3,919
Margin Before Marketing
$125
$175
$255
$383
$588
Marketing
$100
$150
$230
$358
$563
Profit
$25
$25
$25
$25
$25
Margin (%)
15%
15%
15%
15%
15%
Marketing/Sales
12%
13%
14%
14%
14%
ROS
3.0%
2.1%
1.5%
1.0%
0.6%
Year on Year Revenue Growth
—
40%
46%
50%
53%
CAGR Revenue from Year 1
—
40%
43%
45%
47%
Invested Capital
$500
$520
$552
$603
$685
ROI
5.0%
4.8%
4.8%
4.1%
3.6%
Year 5
Cruise
All $ in (Thousands)
Year 1
Year 2
Year 3
Year 4
Year 5
Revenue
$1,320
$1,385
$1,463
$1,557
$1,670
Margin Before Marketing
$198
$208
$219
$234
$251
Marketing
$173
$183
$194
$209
$226
Profit
$25
$25
$25
$25
$25
Margin (%)
15%
15%
15%
15%
15%
Marketing/Sales
13%
13%
13%
13%
14%
ROS
1.9%
1.8%
1.7%
1.6%
1.5%
Year on Year Revenue Growth
—
5%
6%
6%
7%
CAGR Revenue from Year 1
—
5%
5%
6%
6%
Invested Capital
$500
$501
$503
$505
$507
ROI
5.0%
5.0%
5.0%
5.0%
4.9%
384
MARKETING METRICS
On the basis of the information in Table 12.1, most people chose Cruise. Cruise is doing
more with less. It’s more efficient. Its trend in ROS looks much better, and Cruise has
maintained a fairly consistent ROI of about 5%. About the only thing Boom has going
for it is size and growth of the “top line” (sales revenue). Let’s look deeper at the marketing metrics x-ray.
Using the Marketing Metrics X-Ray
Table 12.2 presents the results of our marketing metrics x-ray of Boom and Cruise.
It shows the number of customers each firm is serving and separates these into “old”
(existing customers) and “new” customers.
This table allows us to see not only the rate at which the firm acquired new customers
but also their retention (loyalty) rates. Now, Boom’s spending on marketing looks a
lot better because we now know that spending was used to generate new customers
and keep old ones. In addition, Boom acquires new customers at a lower cost than
Cruise. And although Cruise’s customers spend more, Boom’s stay around longer.
Perhaps we should order another set of x-rays to examine customer profitability and
lifetime value?
Table 12.3 uses the information in the previous table to calculate some additional
customer metrics. Under an assumption of constant margins and retention rates and a
15% discount rate, we can calculate the customer lifetime value (CLV) for the customers of each firm and compare this CLV with what the firms are spending to acquire
the customers. The CLV represents the discounted margins a firm will earn from its
customers over their life buying from the firm. Refer to Section 5.3 for details about
the estimation of CLV and the process for using the number to value the customer
base as an asset. The asset value is merely the number of ending customers times their
remaining lifetime value (CLV minus the just-received margin). For these examples,
we have assumed that all marketing is used to acquire new customers, so the customer
acquisition cost is obtained by dividing marketing spending by the new customers in
year period.
Boom’s aggressive marketing spending looks even better in this light. The difference
between the CLV and acquisition cost is only $3.71 for Cruise but is $48.21 for Boom.
From the viewpoint of the customer asset value at the end of year five, Boom is worth
almost five times as much as Cruise.
Table 12.4 gives us even more information on customers. Customer satisfaction is
much higher for Boom, and Boom’s customers are more willing to recommend the firm
to others. As a consequence, we might expect Boom’s acquisition costs to decline in
the future. In fact, with such a stable and satisfied customer base, we could expect that
brand equity (refer to Section 4.4) measures would be higher too.
Chapter 12 The Marketing Metrics X-Ray and Testing
385
386
MARKETING METRICS
Table 12.2
Marketing Metrics
Boom
Cruise
Year 1
Year 2
Year 3
Year 4
Year 5
Year 1
Year 2
Year 3
Year 4
Year 5
New Customers (Thousands)
1.33
2.00
3.07
4.77
7.50
1.86
1.97
2.09
2.24
2.43
Total Customers (Thousands)
3.33
4.67
6.80
10.21
15.67
3.86
4.05
4.28
4.55
4.88
Sales/Customer
$250
$250
$250
$250
$250
$342
$342
$342
$342
$342
Marketing/New Customer
$75
$75
$75
$75
$75
$93
$93
$93
$93
$93
—
80%
80%
80%
80%
—
54%
54%
54%
54%
Retention Rate
Table 12.3
Customer Profitability
Customer Value Metric
Boom
Cruise
Customer CLV
$123.21
$96.71
Customer Acquisition Cost
$75.00
$93.00
Customer Count (Thousands)
15.67
4.88
$1,344
$222
Customer Asset Value (Thousands)
Chapter 12 The Marketing Metrics X-Ray and Testing
Table 12.4
Customer Attitudes and Awareness
Cruise
Boom
Year 1
Year 2
Year 3
Year 4
Year 5
Year 1
Year 2
Year 3
Year 4
Year 5
Awareness
30%
32%
.31%
31%
33%
20%
22%
22%
23%
23%
Top of Mind
17%
18%
20%
19%
20%
12%
12%
11%
11%
10%
Satisfaction
85%
86%
86%
87%
88%
50%
52%
52%
51%
53%
Willingness to Recommend
65%
66%
68%
67%
69%
42%
43%
42%
40%
39%
387
Hiding Problems in the Marketing Baggage?
The income statement for another example firm, Prestige Luggage, is depicted in
Table 12.5. The company seems to be doing quite well. Unit and dollar sales are growing
rapidly. Margins before marketing are stable and quite robust. Marketing spending and
marketing to sales ratios are growing, but so is the bottom line. So what is not to like?
Table 12.5 Prestige Luggage Income
Statement
Year 1
Year 2
Year 3
Year 4
$14,360
$18,320
$23,500
$30,100
85
115
159
213
Market Share (Unit)
14%
17%
21%
26%
Gross Margin
53%
53%
52%
52%
Marketing
$1,600
$2,143
$2,769
$3,755
Profit
$4,011
$5,317
$7,051
$9,227
ROS
27.9%
29.0%
30.0%
30.7%
Marketing/Sales
11.1%
11.7%
11.8%
12.5%
Sales Revenue (Thousands)
Unit Sales (Thousands)
Using the Marketing Metrics X-Ray
Let’s take a deeper look at what’s going on with Prestige Luggage by examining their
retail customers. When we do, we’ll get a better view of the marketing mechanics that
underlie the seemingly pleasant financials in Table 12.5.
Table 12.6 (refer to Section 6.6 for distribution measures) shows that Prestige Luggage’s
sales growth comes from two sources: an expanding number of outlets stocking the
brand and an increase (more than four-fold) in price promotions. Still, there are plenty
of outlets that do not stock the brand. So there may be room to grow.
Table 12.7 reveals that although the overall sales are increasing, they are not keeping
pace with the number of stores stocking the brand. (Sales per retail store are already
declining.) Also, the promotional pricing by the manufacturer seems to be encouraging individual stores’ inventories to grow. Soon, retailers may become irritated that the
GMROII (gross margin return on inventory investment) has declined considerably.
Future sales may continue to slow further and put pressure on retail margins. If retailer
dissatisfaction causes some retailers to drop the brand from their assortment, manufacturer sales will decline precipitously.
388
MARKETING METRICS
Table 12.6 Prestige Luggage Marketing and Channel Metrics
Year 1
Year 2
Year 3
Year 4
$24,384
$27,577
$33,067
$44,254
87
103
132
183
300
450
650
900
30.0%
22.3%
15.1%
8.9%
ACV Distribution2
30%
40%
48%
60%
% Sales on Deal
10%
13%
20%
38%
Advertising Spending (Thousands)
$700
$693
$707
$721
Promotion Spending (Thousands)
$500
$750
$1,163
$2,034
Retail Dollar Sales (Thousands)
Retail Unit Sales (Thousands)
Number Stocking Outlets
Price Premium
Table 12.7 Luggage Manufacturer Retail Profitability Metrics
Year 1
Year 2
Year 3
Year 4
Retail Margin $
$9,754
$11,169
$13,557
$18,366
Retail Margin %
40%
41%
41%
42%
Retail Inventory (Thousands)
15
27
54
84
Inventory Per Store
50
60
83
93
Sales/Outlet (Thousands)
$81
$61
$51
$49
Stores per Point of AVC %
10
11
14
15
385%
260%
170%
155%
GMROII
In addition, the broadening of distribution and the increase of sales on deal suggest a
possible change in how potential consumers view the previously exclusive image of the
Prestige Luggage brand. The firm might want to order another set of x-rays to see if and
how consumer attitudes about the brand have changed. Again, if these changes are by
design, then maybe Prestige Luggage is okay. If not, then Prestige Luggage should be
worried that its established strategy is falling apart. Add that to the possibility that some
retailers are using deep discounts to unload inventory after they’ve dropped the brand,
and suddenly Prestige Luggage faces a vicious cycle from which it may never recover.
Some things you can’t make up, and this example is one. The actual company was
“pumped up” through a series of price promotions, distribution was expanded, and
sales grew rapidly. Shortly after being bought by another company looking to add to its
luxury goods portfolio of brands, the strategy unraveled. Many stores dropped the line,
and it took years to rebuild the brand and sales.
Chapter 12 The Marketing Metrics X-Ray and Testing
389
These two examples illustrate the importance of digging behind the financial statements using tools such as the marketing x-ray. More numbers, in and of themselves, are
only part of the answer. The ability to see patterns and meaning behind the numbers is
even more important.
Smoking More But Enjoying It Less?
Table 12.8 displays marketing metrics reported by a major consumer-products company aimed at analyzing the trends in competition by lower-priced discount brands.
A declining market size, stagnant company market share, and a growing share of firm
sales accounted for by discount brands all made up a baleful picture of the future. The
firm was replacing premium sales with discount brand sales. To top it off, the advertising and promotion budgets had almost doubled. In the words of Erv Shames, Darden
Professor, it would be easy to conclude that the marketers had “run out of ideas” and
were resorting to the bluntest of instruments: price.
Table 12.8 Market Trends for Discount Brands and Spending; Big Tobacco Company
Year
1987
1992
Market Size (Units)
4,000
3,850
Company Unit Share
25%
24%
Unit Sales
1000
924
Premium Brand Units
925
774
Discount Brand Units
75
150
$600
$1,225
Advertising & Promotion Spend
Table 12.9 Additional Metrics
Year
1987
1992
$1,455
$2,237
Average Unit Price
$1.46
$2.42
Average Premium Price
$1.50
$2.60
Average Discount Price
$0.90
$1.50
$355
$550
Revenue (Thousands)
Operating Profit (Thousands)
390
MARKETING METRICS
The picture looks much brighter, however, after examining the metrics in Table 12.9. It
turns out that in the same five years during which discount brands had become more
prominent, sales revenue and operating income had both grown by over 50%. The
reason is clear: Prices had almost doubled, even though a large portion of these price
increases had been “discounted back” through promotions. Overall, the net impact was
positive on the firm’s bottom line.
Now you might be thinking that the messages in Table 12.9 are so obvious that no one
would ever find the metrics in Table 12.8 to be as troubling as we made them out to be. In
fact, our experience in teaching a case that contains all these metrics is that experienced
marketers from all over the world tend to focus on the metrics in Table 12.8 and pay little
or no attention to the additional metrics—even when given the same level of prominence.
The situation described by the two tables is a close approximation to the actual market
conditions just before the now-famous “Marlboro Friday.” Top management took action
because they were concerned that the series of price increases that led to the attractive
financials in 1992 would not be sustainable because the higher premium prices gave
competitive discount brands more latitude to cut prices. On what later became known as
“Marlboro Friday,” the second of April 1993, Phillip Morris cut Marlboro prices by $0.40
a pack, reducing operating earnings by almost 40%. The stock price tumbled by 25%.
Note in this example the contrast from the preceding example. Prestige Luggage was
increasing promotion expenditures to expand distribution. Prices were falling while
promotion, or sales on deal, were increasing—an ominous sign. With Marlboro, they
were constantly raising the price and then discounting back—a very different strategy.
Marketing Dashboards
The presentation of metrics in the form of management “dashboards” has received a
substantial amount of attention in the last several years. The basic notion seems to be
that the manner of presenting complex data can influence management’s ability to recognize key patterns and trends. Would a dashboard, a graphical depiction of the same
information, make it easier for managers to pick up the ominous trends?
The metaphor of an automobile dashboard is appropriate because there are numerous
metrics that could be used to measure a car’s operation. The dashboard is to provide a
reduced set of the vital measures in a form that is easy for the operator to interpret and use.
Unfortunately, although all automobiles have the same key metrics, it is not as universal across all businesses. The set of appropriate and critical measures may differ across
businesses.
Figure 12.1 presents a dashboard of five critical measures over time. It reveals
strong sales growth while maintaining margins even though selling less expensive
items. Disturbingly, however, the returns for the retailer (GMROII) have fallen precipitously while store inventories have grown. Sales per store have similarly dropped.
Chapter 12 The Marketing Metrics X-Ray and Testing
391
Revenue and Margins
60%
58%
56%
54%
52%
50%
48%
46%
44%
42%
40%
$30,000
$25,000
$20,000
X
X
Year 1
Year 2
X
X
Year 3
Year 4
$15,000
$10,000
$5,000
$0
Gross Margin
Revenue (thousands)
Prestige Luggage Headline Financials
$35,000
The financial metrics look healthy; revenue showing good growth while margins are almost unchanged.
Manufacturer Prices to Store Prices
Store Inventory and GMROII
Prestige Luggage Prices and Retail Prices
$330
Average Price
$280
Q
Avg. Retail
Q Price
Q
Q
$230
$180
X
X
$130
$80
Year 1
Avg. X
Prestige
Price
Year 2
X
Year 3
450%
400%
X
350%
300%
X
250%
200%
X
X
150%
100%
50%
0%
Year 1
Year 4
Prestige Luggage is selling less expensive items.
Distribution
Inventory Per Store
X GMROII
100
90
80
70
60
50
40
30
20
10
0
Year 2
Year 3
Year 4
Prestige Luggage is making diminishing returns for retailer.
Pricing and Promotions
Prestige Luggage Store Distribution
Prestige Luggage Pricing and Promotion
35%
$90
$70
Year 2
$60
Year 3
$50
$40
Year 4
25%
XYear 2
20%
XYear 3
15%
Year 4
X
10%
5%
0%
$30
0%
Sales per Store
We are moving into smaller stores.
Figure 12.1
392
XYear 1
30%
Year 1
Price Premium
$80
10%
20%
30%
% on Deal
Prestige Luggage is becoming reliant on promotion.
Prestige Luggage: Marketing Management Dashboard
MARKETING METRICS
40%
The price premium that Prestige Luggage can command has fallen, and more of the
company’s sales are on deal. This should be a foreboding picture for the company and
should raise concerns about the ability to maintain distribution.
Summary: Marketing Metrics + Financial Metrics = Deeper Insight
Dashboards, scorecards, and what we have termed “x-rays” are collections of marketing
and financial metrics that management believes are important indicators of business
health. Dashboards are designed to provide depth of marketing understanding concerning the business. There are many specific metrics that may be considered important, or
even critical, in any given marketing context. We do not believe it is generally possible
to provide unambiguous advice on which metrics are most important or which management decisions are contingent on the values and trends in certain metrics. These
recommendations would have to be of the “if, then” form, such as “If relative share is
greater than 1.0 and market growth is higher than change in GDP, then invest more in
advertising.” Although such advice might be valuable under many circumstances, our
aims were more modest—simply to provide a resource for marketers to achieve a deeper
understanding of the diversity of metrics that exist.
Our examples, Boom versus Cruise, Prestige Luggage, and Big Tobacco, showed how
selected marketing metrics could give deeper insights into the financial future of companies. In situations such as these, it is important that a full array of marketing and financial metrics inform the decision. Examining a complete set of x-rays does not necessarily
make the decisions any easier (the Big Tobacco example is debated by knowledgeable
industry observers to this day!), but it does help ensure a more comprehensive diagnosis.
12.2 The Value of Information
How much should you spend on gaining information, e.g., market research? Imagine
a firm has three potential marketing strategies: 1) Bold, 2) Moderate, and 3) Cautious.
There are three possible moods that, collectively, the target consumers are in: excited
(40% chance), happy (40% chance), or cynical (20% chance). The firm has to decide
how much to spend to learn the mood of the target consumers.
The cautious strategy will earn $2m in profit whatever mood the consumers are in. The
bold strategy will resonate with excited consumers (earning $10m), perform decently
with happy consumers (earning $2m), but alienate cynical consumers (losing $8m).
The moderate strategy does pretty well with excited ($5m) and happy consumers
($3m), plus it only loses $1m when paired with cynical consumers.
Given this, what is the value of perfect information about the mood of the target
consumers?
Chapter 12 The Marketing Metrics X-Ray and Testing
393
First, calculate how much we can expect to gain with no further information. This is
the maximum of the expected values of the three strategies; where expected value is the
probability weighted average of the outcome values.
Expected Value, Bold: 40% × $10m + 40% × $2m + 20% × -$8m = $3.2 million
Expected Value, Moderate: 40% × $5m + 40% × $3m + 20% × -$1m = $3 million
Expected Value, Cautious: Whatever the consumers’ mood = $2 million
Thus, without any additional information on the consumers’ mood we’d choose the
Bold strategy as its expected value ($3.2 million) is the highest.
If market research could give us perfect information this would allow us to pick the
best strategy to pair with the consumers’ mood, i.e. Bold with Excited (gaining $10m),
Moderate with Happy (gaining $3m), and Cautious with Cynical (gaining $2m). With
perfect information we would expect a profit of:
Expected Value with Perfect Information
40% × $10m + 40% × $3m + 20% × $2m = $5.6m
Because the expected value with perfect information is $5.6 million and the expected
value without additional information is $3.2 million, the expected value of the perfect
information is the difference between the two or $2.4 million.
This quantity is an upper bound on the value of any actual information the firm can
collect. In the real world the value of any imperfect information the firm can collect
must be less than this upper bound.
These calculations assume marketers only care about expected value when, in fact, risk
is also a concern. Firms prefer a certain $10 million to a 50% chance of gaining $20
million and 50% chance of gaining nothing even though the expected values are the
same. This is known as risk aversion. If you are risk averse you may wish to pay for
information that reduces the range of outcomes that you face even if this doesn’t change
the expected value—a consideration not taken into account in the calculation of the
expected value of perfect information.
Individual decision makers are also often loss averse. When loss averse, decision makers are willing to reduce the expected value of a decision in order to limit potential
losses. Unlike risk aversion, loss aversion is often viewed by economists as poor decision making. A marketer who knows they will be sacked if they lose money in the above
scenario might select the cautious strategy which, although it has the lowest expected
value, never loses money. The actual value of perfect information to risk averse and
loss averse decision makers is usually higher than expected value of perfect information
because the information not only improves expected value but decreases risk/loss.
394
MARKETING METRICS
In summary, the value of information and so the usefulness of market research and testing varies with the precise situation at hand. Since reality is decidedly more complicated
than our illustrative example, allocating data collection and analytical resources is an
important management decision. Estimating the value of information requires specific
quantitative inputs, which are often assumptions. Unfortunately managers may be sufficiently unsure of these inputs that they are unable or unwilling to quantify these estimates. Even in these instances, however, it may be worthwhile to develop an intuitive
appreciation of when additional information is likely to be most valuable. Table 12.10
should be useful in making qualitative comparisons of situations in which managers are
uncertain about the value of collecting further data to refine their choices.
Table 12.10
Quick Guide to the Value of Information
Criteria
Information Most Valuable
When
Information Least Valuable When
Potential financial
consequences of
decision
Large difference between the
consequences of the best and
worst alternatives
Small differences between the
consequences of the best and worst
alternatives
Uncertainty of future High degree of uncertainty
outcomes
Low degree of uncertainty
Ability of information Information is likely to change
to change decision
the decision. (A combination of
powerful information and close
initial decision)
Information is not likely to change
the decision (A combination of
poor information quality and
obvious initial decision)
Validity of metrics
Metrics are valid indicators of
market outcome
Metrics are biased indicators of
market outcomes
Reliability of data
When the sample size is large
and measurement error is small
The sample size is small and
measurement error is large
12.3 Testing
Testing usually underpins successful marketing. When you are not sure which advertising copy to use, which marketing mix elements to emphasize, or even which product
variants to offer, testing can help. When testing, you should consider the precision
(known as reliability) of the test. Testing an advertisement on one person will be unreliable as each person has idiosyncrasies. Increasing the number of respondents in the test
increases your confidence that the responses are typical of the group being tested. You
must also consider the validity of your test; e.g., are you testing the right group? Are you
Chapter 12 The Marketing Metrics X-Ray and Testing
395
asking the right question? If you test an advertisement on bank managers, this may not
be valid for estimating how your target market, college athletes, will react.
We have already mentioned side-by-side A/B testing, where two versions of an advertisement are created and tested against each other. Online it is easy to serve randomly
selected visitors different versions of the advertisement. Randomly selecting which
visitors get which advertisement ensures that there is no systematic difference between
who gets which advertisement suggesting that any observed difference is driven by the
advertisement. Online it is usually relatively cheap to create and test new versions of
advertisements. That said even online testing is not free and in general the cost of any
testing can be high.
The more versions of an advertisement we create and test, the greater the chance of
finding excellent copy. Unfortunately, creating and testing versions reduces the money
available to spend on deploying whichever version wins the test. The Gross model is
designed to help managers make this tradeoff.
The Gross Model
The Gross model, named after Irwin Gross, advises how much of the budget should
be spent on creating (and testing) advertising copy. The number of alternative copies
you should be willing to develop depends upon the variability of the effectiveness of
the advertisements. If some advertisements are highly effective, but most ineffective,
you will want to spend relatively heavily on developing and testing copy. The potential
upside is high and you want to develop many versions in order to get a great version.
If, however, all advertisements perform relatively similarly, the difference between what
you already have, and what you can gain with further development, is quite limited.
In that case it is better to spend less on developing new copies and spend more of the
budget showing the currently best advertisement.
In the Gross model, the effectiveness of an advertising campaign (Z) is the amount
spent on buying media (D), multiplied by the effectiveness of the best advertisement
created, (E). So Z = E*D
If one assumes that advertising effectiveness is linear, i.e., each piece of spending is
equally effective, this is a simple model. Unfortunately advertising often takes multiple
views to gain any traction and eventually loses effectiveness: the S curve described in
Chapter 9. This means E varies with the amount spent on media (D) making the model
more difficult to use.
The amount spent on media (D) is the total budget (B) less the fixed costs of copy testing (CF) and the total costs of each new advertisement. This is the number of advertisements (N) developed multiplied by the average cost to develop an advertisement (C)
plus the marginal (extra) costs to test each advertisement (CS).
396
MARKETING METRICS
So D = B – CF – N(C+CS)
O’Connor and her colleagues in 1996 applied this formula using the historical distribution of advertising effectiveness and concluded that 20-30% of media budgets (B)
should be spent on developing and testing, with the rest (D) spent on deployment. Of
course this rule of thumb may vary significantly depending on context.
Should You Test Another Advertisement?
The original Gross model was designed to help managers decide how many advertisements to create and test. The general conclusion was that firms tended to spend too
little on creating and testing alternative versions. (Perhaps because this budget item is
often called “non-working” media expense).
We will ask a slightly different question. Should the firm develop an additional advertising copy execution or use all of the remaining budget to air the current best copy?
Let B be the total budget to create, test, and deploy the best testing advertising. Let C
be the cost to develop and test a new version of the advertisement, i.e., the expenditure
incurred when you commission a new piece of copy. Because tests are never perfectly
reliable or valid, we propose a “vaguely right” adjustment factor (A). The formula for
the adjustment factor is
1
(an approach similar to that proposed by
Reliability * Validity
Irwin Gross). As the reliability and validity of the tests approach 1 the adjustment factor
approaches 1. As A gets higher, i.e., whenever the validity and reliability of the test gets
lower, each test is less useful at predicting real-world performance. The intuition is that
with low reliability and validity, the test would need to indicate a higher probability of
increased effectiveness before we would spend the money for the test. In our model we
use A=2, which means the test would need to indicate an expected return 2X the cost
before we would deem it acceptable. This 2X would, for example, result from reliability
of 70% and validity of 70%.
We can now estimate how much lift we need to expect to gain from a new version of
an advertisement to make commissioning it worthwhile. This breakeven lift is the cost
of commissioning the new version as a percentage of your free budget multiplied by
the adjustment factor. If we have a budget of $2.25 million and the cost to develop and
test each version is $40,000, then developing a new version costs 1.8% of the budget. If
the adjustment factor is 2 we must expect to gain at least 2*1.8% = 3.6% performance
lift to make commissioning a new version of the advertisement worthwhile; see Table
12.11. Of course, this approach means that each time we spend money for a copy test
the remaining budget is also reduced and the next copy decision will represent a higher
percentage of the budget and, therefore, require a higher percent expected increase in
sales to justify undertaking the test.
Chapter 12 The Marketing Metrics X-Ray and Testing
397
Table 12.11 (Breakeven) Lift Needed to Commission New Piece of Copy
Free Budget to Buy Media or Develop and Test Versions (B)
$2,250,000
Cost to Develop and Test New Version (C)
$40,000
Developing and Testing as % of Budget (D=C/B)
1.8%
Adjustment Factor for Lack of Test Reliability and Validity (A)
2
Expected Lift That Justifies Testing New Version (JT=A*D)
3.6%
Will a new version of the advertisement give sufficient lift to make creating it worthwhile? The challenge in answering this is that we don’t know any version’s effectiveness before it has been created and tested; i.e., after we have already spent the money
developing it.
The expected value of the new version is the chance of getting a new version of a certain
quality multiplied by the outcome when we get a new version of that quality. There are
two broad outcomes. The new version is equal or of worse quality than our best current advertisement. When this is the case the new version has no value at all. When the
new version is of higher quality then lift is the additional value coming from the new
version.
We make the assumption that each version of an advertisement has a quality score
between one and ten and a quality of 10 is twice as effective—in some way defined by
the firm—as 5, etc. For our example we assume that there is a uniform distribution of
quality for versions of the advertisement. If we have ten levels of quality (1-10) each is
equally likely. (Note one of the advantages of this model is that you can specify any distribution of advertisement quality that you wish, just change the distribution in Table
12.12.) The lift from the new version is the quality of the new version minus quality of
the current version divided by the quality of the current version. We know the quality of our current version—what we will deploy if we end testing—and so can create
an expected lift from a new version. If this exceeds the expected lift needed to justify
commissioning a new version, Table 12.11, we should do so, otherwise we should stop
testing and deploy our current version. Table 12.12 shows us that with a current version
quality 8 we should continue testing, but it is a very close call. (To read Table 12.12,
note that the current advertisements scores an 8. Looking across shows us that a new
version that scores 8 or less will not provide an effectiveness lift. A new advertisement
that generates a score of 9 will improve life by 12.5%.)
398
MARKETING METRICS
Table 12.12 Expected Lift From New Advertisement
Quality of Current Version (QCur)
8
Quality New
Version (QNew)
Chance of
Occurring (CO)
Lift From New
Version (LV)**
LV × CO
1
10%
0
0
2
10%
0
0
3
10%
0
0
4
10%
0
0
5
10%
0
0
6
10%
0
0
7
10%
0
0
8
10%
0
0
9
10%
12.5%
1.25%
10
10%
25.0%
2.5%
Expected Lift of New Version (ENew)
3.75%
** Lift From New Version =IF(QNew>QCur, (QNew-QCur)/QCur, 0)
Whether you should continue creating versions depends on the quality of your current
version. Table 12.13 shows the expected lift at each quality of current version. When
you only have a low quality current version, developing new copy has a huge expected
value as you are likely to get copy that will substantially increase the effectiveness of
your media spending. You should stop testing if you have a current version of quality
9—the expected benefits of continuing to commission versions are less than the cost
of doing so. If you already have a version of quality 10 there is no possible benefit to
further testing.
Chapter 12 The Marketing Metrics X-Ray and Testing
399
Table 12.13
Expected Lift From New Version Given Current Version Score
Quality Current
Version (QCur)
Expected Lift From
New Version (ENew)
Decision (Test if ENew>JT)
1
450.0%
Test As > 3.6%
2
180.0%
Test As > 3.6%
3
93.3%
Test As > 3.6%
4
52.5%
Test As > 3.6%
5
30.0%
Test As > 3.6%
6
16.7%
Test As > 3.6%
7
8.6%
Test As > 3.6%
8
3.75%
Test As > 3.6%
9
1.1%
Don’t Test as < 3.6%
10
0.0%
Don’t Test as < 3.6%
This model contains a number of assumptions, such as that the version of an advertisement can be neatly scored and the quality translates predictably into relative sales
results. Furthermore, we also assume each new copy comes from the same distribution;
i.e., this somewhat implies the agency isn’t giving you its best ideas first. This need
not be true. You might get progressively worse ideas each time you go to the agency.
Despite the challenges in getting a perfect model, we think the method is most valuable
as a general illustration of how to think about the value of information in a management decision context that includes uncertainty, financial consequences, and ability of
imperfect data to inform a decision.
References and Suggested Further Reading
Ambler, Tim, Flora Kokkinaki, and Stefano Puntonni (2004). “Assessing Marketing
Performance: Reason for Metric Selection,” Journal of Marketing Management, 20,
pp. 475–498.
McGovern, Gail, David Court, John A. Quelch, and Blair Crawford (2004). “Bringing
Customers into the Boardroom,” Harvard Business Review, November, pp. 1–10.
Meyer, C. (1994). “How the Right Measures Help Teams Excel,” Harvard Business Review.
72(3), 95.
O’Conner, Fina Colarelli, Thomas R. Willemain, and James MacLachlan. (1996) “The Value
of Competition among Agencies in Developing Ad Campaigns: Revisiting Gross’s Model.”
Journal of Advertising 25 (1), pages 51–62.
400
MARKETING METRICS
13
SYSTEM OF METRICS
“There are three kinds of economists: those who can count and those who can’t.”
—Unknown source
Modeling Firm Performance
To better understand the factors contributing to overall firm success, managers and
analysts often decompose return on assets (ROA) into the product of two ratios
with each ratio reflecting a different aspect of the business. One popular approach or
“model” for this decomposition is the DuPont Model.
ROA =
Net Profit
Sales
×
Sales
Assets
The first ratio in this simplified DuPont Model is called either the profit margin or
return on sales. It measures how profitable is each dollar of sales. To the extent that
marketers create products that customers value, claim that value through intelligent
pricing, drive down costs by paying attention to manufacturing and channel costs, and
optimize their marketing spending, marketers can increase the firm’s return on sales.
The second ratio in the DuPont Model is known as asset turnover. Asset turnover can
be thought of as the number of dollars of sales each dollar of assets generates. Here the
job of marketers is even more focused—on generating dollars of sales but with an eye
toward managing assets such as inventory and receivables captured in the denominator.
Notice that the DuPont Model is an identity.1 It is always true regardless of the values
taken on by the various ratios. It is always true mostly because we have defined the
ratios in such a way so as to make it always true. So it makes no sense to argue with or
take exception to the DuPont Model.
But if it is simply an equation that is true by definition, what good is it?
401
It is useful to the extent that the decomposition of ROA into the two component ratios
helps firms maximize ROA by focusing (separately) on the two components. It is also
useful in that it reminds marketers that their job is not simply to generate sales, but to
generate profitable sales and to do so efficiently (with respect to assets used).
The DuPont Model has demonstrated its usefulness in practice. A Google search resulted in 4.4 million results for “DuPont Model” compared to 2.9 million results for
“DuPont Chemicals.” In some circles, the company is now more famous for its model
than its chemicals.
Figure 13.1 illustrates how the DuPont Model is often expanded to include components
affecting the two input ratios.
The DuPont Model
Cost of Goods
Sold
Measures of the
effectiveness with
which assets are
used to produce
revenues.
S, G & A
Expenses
Interest
Expense
Sales
Total
Costs
Net
Profit
-/-
Income
Taxes
Sales
Net Profit
to Sales
Cash
Measures of
investments in
working capital assets
needed for sustaining
ongoing operations.
x
Accounts
Receivable
Inventories
Return on
Assets
Current
Assets
Marketable
Securities
Sales
Sales to
Assets
-/-
Other
+
Total
Assets
Land
Measures of
investments in longterm, revenueproducing assets.
Buildings
Machinery and
Equipment
Non-Current
Assets
Intangibles
Figure 13.1 An Extended DuPont Model (adapted from http://www.12manage.com/
methods_dupont_model.html)
Notice that the three rightmost columns of boxes in Figure 13.1 represent the DuPont
Model. The two leftmost columns of boxes represent a particular method of breaking
apart net profit and total assets into smaller components. Our purpose here is not to
402
MARKETING METRICS
critique the above representation of the components of firm performance, but simply
to offer a few observations. First, we note that the decompositions of total costs, current assets, and non-current assets should be familiar to most readers. The categories
of components used are consistent with what one finds on income statements (where
total cost appears) and balance sheets (where total assets appears). Second, we note that
the assets that marketing creates (brands and customer relationships, for example) get
lumped together as intangibles signaling that they are difficult to measure (which we
agree with) and perhaps an afterthought or “other” category (which we disagree with).
Finally, and most importantly, we observe that although total cost, current assets, and
non-current assets all get broken out into smaller, well-understood components, sales
does not. It is as if Costs and Assets deserve a lot of attention but the components of
sales do not. This is perhaps not surprising since this particular model was designed
by finance and accounting executives. As marketers, however, much of our focus is on
how sales are generated. We also care about costs and asset utilization, of course, but we
care more about sales and the components of sales. Figure 13.1 reflects the inward focus
of a firm whose success depended on making things, minimizing costs, and using assets
efficiently. For today’s firms whose success depends at least as much on marketing and
sales as production, we need a different model. We need our own “DuPont Model,”
with at least the same amount of detail and clarity for breaking down the components
of sales as the commonly used breakdowns of costs and assets.
Of course, as we begin to think about how to break down sales into its components, we
quickly come to understand why there is no commonly used breakdown across different types of businesses. As all marketers know, there are multiple ways to decompose
or break down sales simply because several entities (most of them outside the firm) are
involved in the creation of revenue: sales force, customers, dealers, and even our competition. With a multitude of insightful ways to break down sales, it is no wonder there
is not one commonly accepted way.
To illustrate, Figure 13.2 shows four (of many) separate and valid ways to break down
sales into smaller components.
■
SALES = Number Salespersons * Avg. Sales/Salesperson
■
SALES = Number dealers ACV% * Avg. Sales per dealer ACV%
■
SALES = Our Dollar Share * Total Market Sales
■
SALES = Number customers * Sales per customer
As with the DuPont Model, each of four ways to compute sales is an identity. Sales will
always equal the number of customers times the average sales per customer. But even
though they are identities, they can still lead to valuable insights as we will attempt to
demonstrate.
Chapter 13 System of Metrics
403
Number
Dealers
ACV%
Number
Salespersons
Avg. Sales
per Dealer
ACV%
Average
Sales/Salesperson
Sales
Our
Dollar
Share
Number
Customers
Sales per
Customer
Market
Sales
Figure 13.2 A Sales Model
We also point out that there will be other ways to break down sales. Figure 13.2 simply
illustrates four ways. Also know that the outer ring of components of sales in Figure
13.2 can themselves be decomposed. For example: Sales per customer can be calculated
as Purchases per customer (per period) * Average Sales per purchase. And, not unexpectedly, there will be multiple ways to decompose each of the outer-ring components.
For example Sales per customer can also be decomposed into Units purchased per customer * Average price per unit. Decomposing the components of sales can be thought
of as expanding the diagram in Figure 13.2 outward. We might also think of expanding
the model “upward” with separate pages (decompositions) for each product or each
customer group or each vendor.
Three Reasons for Using Systems of Identities in Marketing
There are three primary reasons for formulating marketing DuPont-like component
models of your marketing decisions and objectives:
1. Decomposing the metric of interest into components can make it possible
to identify problems and opportunities for improvement in more detail. For
example, did share drop because our sales were down or competitors’ sales were
up? If our sales were down, was that due to fewer customers buying, lower unit
sales per customer, lower average prices, or some combination? Decomposition
may also help by separating identities from empirical relationships. Although
404
MARKETING METRICS
identities are easy (just arithmetic), empirical relationships require difficult
judgments about the form of the relationship, causality, and the future.
2. Decomposing metrics may also allow us to estimate, indirectly, other component metrics that are difficult to measure directly. Using multiple identities can
help eliminate measurement error with multiple “checks” on the value of any
specific metrics. In the same way, individual marketing metrics may be regarded as part of a network or “web” of relationships. If each link in the network is
valid, even if individual values are estimated with error, the entire structure will
be more robust.
3. Selecting and organizing the right network of marketing metrics often helps
formulate models of marketing mix decisions. Like the DuPont Model, using
models with interim components can make such models and dashboards more
managerially transparent and help managers make and monitor the effects of
their decisions.
Decomposing for Diagnostic Purposes
As mentioned previously, a primary purpose for using one or more identities to decompose any marketing metric of interest is to gain a deeper understanding (or at least a
different perspective) on the reasons for changes and differences observed. Although
identities may be developed with a view to understanding the sources of changes and
differences, they do not require calibration or estimation. They are true by definition,
and we will designate these with an (ID).
An example of an identity is the relationship between Sales, Quantity, and Price:
Sales = Quantity × Price (ID)
If we witness declining sales, the identity helps us see, first, whether the decline was due
to declining quantity or price or both. And next it helps us understand that if quantity
declined, price increased, but sales declined, that quantity must have declined by a
larger percentage than did the price increase.
In contrast to identities are empirical relationships—relationships between variables
for which the exact equation is not known and/or for which the relationship holds
only imperfectly. Empirical relationships are required, for example, to help us decide
whether we should increase or decrease prices. We designate these with an (EM). For
example, we might consider the relationship between quantity sold to be a direct, linear
function of price charged:
Quantity = b × Price + error (EM)
Chapter 13 System of Metrics
405
This empirical relationship between quantity and price necessarily contains an error
to account for measuring price or quantity imperfectly or influences on quantity sold
other than price (our competitors’ prices, for example). Also note that the parameter
“b” in this empirical relationship is, itself, a variable. It is an unknown constant—one
that we might, for example, be able to estimate from available data. But one of the key
differences between identities (ID) and empirical relationships (EM) is that empirical
relationships are more flexible. They apply to the tough and important questions such
as “how many more units will we sell if we lower the price by $1?”
Dashboards of metrics often reflect underlying management logic about how marketing works to influence sales and profits. Dashboards include both identities and empirical relationships. As illustrated in Figure 13.2, sales can be decomposed many ways.
Some of the components of sales might themselves be decomposed using one or more
identities. Each firm needs to identify its primary performance measures. This is what
should appear on their dashboards. There should be the capability to drill down on
each of these performance measures (using identities) to diagnose and explain changes
across time. But if dashboards are to be more than monitoring devices, we should have
some idea of causal connections (step on the brake to slow down the vehicle, step on the
accelerator to make it go faster). Before long, they can become complicated as we start
to take into consideration the multiple effects of some of the variables, e.g., step on the
accelerator to make the car go faster and the fuel gauge drops. Sometimes we also need
a system of metrics to help infer (or forecast) values that are difficult to measure directly
(e.g., how much farther can we drive before the gas tank is empty?).
Eliminating Error by Harnessing the Law of Large (and Not So Large)
Numbers
There is the classic story of the physics professor whose final exam asked students to
explain how to use a barometer to measure the height of a building. In addition to the
“obvious” answer to measure the barometric pressures at the top and bottom of the
building and use the difference to calculate the building’s height, the professor purportedly received several other creative answers. Drop the barometer from the top of
the building, time how long it takes to hit the ground, and use the appropriate physics
formula to infer the height. Tie the barometer to a string, lower it to the ground, and
measure the length of the string. Measure the length of the shadow cast by the building, the length of the shadow cast by the barometer, the height of the barometer, and
use proportions to calculate the height of the building. By far the most creative solution
purportedly offered was to knock on the door of the building’s janitor and offer to give
the janitor the barometer in exchange for revealing the height of the building.
The multiple ways to calculate sales shown in Figure 13.2 are similar to the multiple
ways students came up with to measure the height of the building. Rather than argue
over which single method to use, we propose to look for a way to use them all. When
406
MARKETING METRICS
faced with a dilemma of which of two methods to use, why not do both? For the
barometer problem, why not use several different methods and then combine the many
estimates into one final estimate—perhaps by doing something as simple as taking
the average of the estimates. If we wanted to do a little bit better, we could calculate
a weighted average with weights depending on some measure of how “accurate” each
estimate was. We might put more weight on the string-based estimate and less on the
estimate from timing the fall of the barometer if we thought our watch and wind made
the timing-based estimate less accurate. The relative weight to put on the janitor’s estimate would depend on our confidence in the estimate. If the janitor claims to “know”
the height, we should give the estimate more weight than if the janitor admits the number is something of a guess.
Using the average of the estimates instead of any one of the estimates takes advantage
of the law of large (and not so large) numbers. The average is expected to be closer to
the true value and become closer the more estimates that we have to average together.
Ideally we want “independent” estimates such as might be the case with the barometer
example (unless, of course, the janitor got his number using the string method).
In the barometer example, we were mostly interested in measuring the height of the
building. In our example, marketers are probably just as interested in the measuring
components as we are in measuring sales itself. In fact, it often is the case that the firm
has a good handle on sales and would like to get a better measure of some of the components such as share or the sales per customer or any of the other metrics in the outer
ring or outer-outer ring. In extreme cases, the firm may have no separate measure of
one of the components and will have to “back into it” based on the measurements of
all the others. (In the barometer example, use the height of the building and the length
of the barometer’s shadow to estimate the length of the building’s shadow—to measure
how far away the building is, for example, without having to travel to the building.)
What this means is that every initial estimate (and the associated standard deviation)
will combine to determine our final estimates. Our estimate of the length of the string
will be used to help revise our estimate of the time it took the barometer to hit the
ground and the length of the building’s shadow and vice versa. We think it is easy to see
that the more separate estimates and identities we have in the model, the more confident we are with the final estimate.
Whereas the carpenter adage is to measure twice and cut once, here we say measure
many times and many ways and put them all together in a systematic, logical way. Use
not only a square to check for a right angle, also measure 3 feet and 4 feet along each
side and check to see if the diagonal measures 5 feet. That’s the idea behind the proposed process for fine tuning a system of marketing metrics.
Chapter 13 System of Metrics
407
Using Identities to Estimate Metrics that are Difficult to Measure Directly
“Decomposition involves figuring out how to compute something very uncertain from other things that are a lot less uncertain or at least easier to measure.”
(Hubbard, 2007)
Marketing models can often make use of our ability to infer missing variables through
construction of the appropriate identity. First, let’s take an example from the physical
world and use that to draw a parallel to marketing problems. If you wanted to calculate
directly the average depth of your local swimming pool, that would involve a series of
complicated and difficult measurements (either measuring the depth repeatedly as one
moved across the length and width of the pool or somehow capturing the curve of the
bottom with a functional form and using calculus and algebra). An indirect method
might be easier. Record the volume of water required to fill the pool and divide by the
pool’s surface area.
Marketers are also often interested in estimating the values that are conceivably directly
measurable, yet might be more efficiently estimated from combinations of other metrics. An example is a firm’s average Share of Requirements or Share of Wallet either
in dollars or in units. To measure this directly would require a database of customer
purchases that included its own firm purchases and all other purchases in the same category. Further, the customers included in the database would need to be representative
of the entire category or weighted in an appropriate way. Instead of a direct measurement, marketers might find it easier and more efficient to estimate share of requirements from the equation included in Sections 2.3 and 2.5:
Share of Requirements (%) =
Market Share (%)
Penetration Share ($,#) * Usage Index ($,#)
The latter three variables might be directly measurable from reported sales, a count of
known customers, and an estimate of the degree to which the firm’s own customers
are heavy or light users of the category. Of course, the metric estimated in this manner
is an average and will not give insight into the variation in customer loyalty behavior
represented by the metric.
Marketing Mix Models—Monitoring Relationships between
Marketing Decisions and Objectives
As Neil Borden, Sr., the author of the term “marketing mix” noted a half-century ago,
“Several characteristics of the marketing environment make it difficult to predict and
control the effect of marketing actions.”2 A system of marketing identities can help with
408
MARKETING METRICS
this problem by providing integrated frameworks and structures for monitoring the
outcomes from marketing decisions. Marketing models must often trade off comprehensiveness with comprehensibility, completeness with simplicity.
The complexities include these: First several potential marketing actions may affect
sales and profits. These potential actions include pricing, price promotion, advertising,
personal selling, and distribution changes, to name just a few. Second, the effects of
any one of these actions on sales, even holding all of the other actions equal, are often
non-linear. The infamous S-curve is an example of this non-linearity (a little advertising produces no effect, somewhat more stimulates sales, and at some point effectiveness
diminishes and disappears altogether). Third, the effects of one marketing decision
often depend upon other marketing decisions. For example, the effects of advertising on
sales depend not only on the product design, but also on price and product availability.
Fourth, there are also “feedback” and lagged effects in marketing. Over time, our investments in advertising might build brand equity that allows our brand to charge higher
prices. Or, if competitors introduce a better product and sales fall to the point that
salespeople are earning too little, the same salespeople may resign or spend less time on
a particular product line, causing sales to fall again. The potential complexity resulting
from specifying a large number of marketing mix elements, non-linearities of effects,
interactions among elements, lagged and feedback effects, and competitive behavior is
mind-numbing. Further, these potential complexities seem to be limited only by the
imagination—and marketing people are (by definition?) creative! It is simply not possible, we assert, to capture all of these complexities with any empirical model.
In the face of such potential for complexity, it is important that marketers find
approaches that will help them, in the words of Arnold Zellner, keep it sophisticatedly
simple (KISS—we know you thought it stood for something else).3 Careful selection
of marketing metrics frameworks that are constructed around a few important identities has several benefits. One is that they enable us to specify the most important
interactions and feedback loops at the level of structural identities instead of empirical
relationships.
Let’s begin by distinguishing between marketing decisions (actions), objectives (for
example, profits), and intervening metrics that help us understand the connections. A
simple marketing mix model might be the following: profits = f(unit price, advertising,
sales force, and trade promotion), which written out in English means profits are a
function of unit price, advertising, sales force, and trade promotion (see Figure 13.3).
Many marketers would reject the model in Figure 13.3 as not sufficiently detailed as
concerns the multiple effects of marketing mix decisions. A $1 increase in Unit Price,
for example, would result in a $1 increase in unit margin while, probably, decreasing
unit sales. Estimating the empirical relationship between unit price and unit sales separately, and then making use of identities involving unit price, unit cost, and unit sales
Chapter 13 System of Metrics
409
to calculate gross profit (as illustrated in Figure 13.4), is generally preferred. Thus, we
separate what can be calculated (using an identity) from what must be estimated (using
an empirical relationship). Similarly, knowing the causal effect of advertising, sales
force, and trade promotion spending on unit sales allows the marketer to calculate the
effect on profits and determine whether an increase or decrease is justified (see Figure
13.4). The usefulness rests on the assumption that we will do a better job of understanding marketing mix effects by separating those that must be empirically estimated from
others that are governed by accounting identities.
Unit Price
Advertising
Net Profit
EM
Sales Force
Trade Promotion
EM Empirical Relationship
ID Identity Relationship
Objectives
Decision
Figure 13.3 Empirical Relationships between Marketing Decisions and Objectives
Unit Cost
Unit
Margin
ID
Unit Price
ID
Gross
Profit
Unit Sales
Advertising
EM
Sales Force
Net Profit
ID
Trade Promotion
Marketing
$
EM
Empirical Relationship
ID
Identity Relationship
Objectives
Metrics
ID
Decision
Figure 13.4 Empirical Relationship with Components of Marketing Outcomes
410
MARKETING METRICS
Marketing mix models are used to estimate the effects of marketing levers on marketing objectives and make decisions about how to allocate resources. One of the most
frequently applied marketing mix models is the one underlying simulated test markets
and depicted in Figure 13.5. With only minor variations, these models are used to forecast new product sales (see Section 4.1 for more detail). The structure of this model is
straightforward, even if some would argue it is not simple. Forecast unit sales are calculated in a multiplicative identity from the metrics below. The multiplicative nature
of the identity captures the most significant interactions of the marketing mix without
resorting to (even more) complex equations. It is, we assert, more managerially transparent and useful because of this well-structured system of metrics that defines and
separates identities from empirical relationships.
Forecast Unit Sales = Number of Consumer Prospects *Awareness * Availability
*( Trial Rate * Trial Units + Repeat Rate * Repeat Units).
The input estimates for the components are obtained from the results of the simulated
test, surveys, management judgment, and/or empirical models.
One of the advantages of the model in Figure 13.5 is that it also provides clear and separate paths by which the different marketing mix elements are believed to impact unit
sales. Advertising affects Consumer Awareness but not Availability. Of course, in reality “everything affects everything,” but the KISS structure affords a transparency and
utility that might be destroyed if management didn’t impose the discipline of focusing
on the most important empirical relationships that the identity relationships suggest.
In the case of the new product forecasting model in Figure 13.5 we have decomposed
(defined) the forecast sales to be a function of the metrics listed. The way we choose
to decompose the objective may be more or less suitable for separating marketing mix
empirical effects. For example, breaking down a share goal into share of requirements,
usage index, and penetration share would not have an obvious relationship to individual mix elements. Everything would still affect everything. So, not every identity will
be helpful in a model of marketing mix effects.
Chapter 13 System of Metrics
411
Advertising
Number of
Consumer
Prospects
Consumer
Awareness
EM
Sales Force
Forecast
Unit Sales
Availability
(ACV%)
EM
Trial Rate%*
Trial Units
EM
Trade Promotion
ID
Product Concept
Unit Price
Repeat Rate%*
Repeat Units
EM
Product Quality
EM
Empirical Relationship
ID
Identity Relationship
Objectives
Metrics
Decision Levers
Figure 13.5
Simulated Test Markets Combine Empirical and Identity Relationships
Also, depending on how the data are collected, some identities may be strongly suggested by the data, even if they are not directly measured. For example, in consumer
packaged goods markets, data on distribution (see Section 6.6) and channel promotion
activity (incremental sales lift %—see Section 8.1) are regularly collected and reported
to marketing managers. The availability of these two metrics strongly suggests the need
for a third metric, “preference,” to create an attractive identity that may be useful in
separating empirical effects and allowing for important interactions. Figure 13.6 shows
how marketers might be able to “back into” values of preference by combining share,
lift %, and distribution metrics. Of course, this approach means that the marketers are
defining preference in a way that is consistent with relative choice under scenarios of
equal distribution and lift %.
412
MARKETING METRICS
Unit Price
Consumer
Preference
%
Advertising
EM
Sales Force
Share
Promotion
Lift %
ID
EM
Trade Promotion
Distribution
PCV %
EM
Empirical Relationship
ID
Identity Relationship
Metrics
EM
Constructs*
Decision Levers
Figure 13.6
Empirical Relationship with Marketing Components
and Intermediate Metrics and Constructs
Related Metrics and Concepts
By definition, accounting identities always hold. It is simply a matter of getting the correct values for the component parts. Other identities, such as those found in theoretical
models of finance and economics, are true “in theory” or assuming certain conditions.
For example, as discussed in Sections 7.3 and 7.4, at profit-maximizing levels of price,
this identity should be true:
Margin on Sales [(Price – Variable Cost)/Price] = 1
price elasticity for constant elasticity demand, or
Price = Variable Cost + 1⁄2 (Maximum Willingness to Pay
– Variable Cost) for linear demand functions
These identities identify relationships that are unlikely to be precise, but are vaguely
right.
Chapter 13 System of Metrics
413
References and Suggested Further Reading
Hubbard, Douglas W. (2007). How to Measure Anything: Finding the Value of “Intangibles” in Business, John Wiley & Sons, Hoboken, New Jersey.
Conclusion
“. . . metrics should be necessary (i.e., the company cannot do without them), precise,
consistent, and sufficient (i.e., comprehensive) for review purposes.” 4
Understanding metrics will allow marketers to choose the right input data to give
them meaningful information. They should be able to pick and choose from a variety
of metrics depending upon the circumstances and create a dashboard of the most vital
metrics to aid them in managing their business. After reading this work, we hope you
agree that no one metric is going to give a full picture. It is only when you can use multiple viewpoints, known as triangulating because you come at the problem from different angles, that you are likely to obtain anything approaching a full picture.
“. . . results measures tell us where we stand in efforts to achieve goals, but not how
we go there or what to do differently”. 5
Marketing metrics are needed to give a complete picture of a business’s health. Financial
metrics focus on dollars and periods of time, telling us how profits, cash, and assets are
changing. However, we also need to understand what is happening with our customers,
products, prices, channels, competitors, and brands.
The interpretation of marketing metrics requires knowledge and judgment. This
book helps give you the knowledge so that you can know more about how metrics are
constructed and what they measure. Knowing the limitations of individual metrics is
important. In our experience, businesses are usually complex, requiring multiple metrics to capture different facets—to tell you what is going on.
Because of this complexity, marketing metrics often raise as many questions as they
answer. Certainly, they rarely provide easy answers about what managers should do.
Having a set of metrics based on a limited, faulty, or outmoded view of the business can
also blind you. Such a set of metrics can falsely reassure you that the business is fine
when in fact trouble is developing. Like the ostrich with his head in the sand, it might be
more comfortable to know less. We recommend you guard against the classic problem
of measuring what you can easily measure regardless of whether this is helpful to your
business. It is better to consider what metrics will be useful for your business and work
to find a way to access them.
414
MARKETING METRICS
We would also recommend you spend time considering how to communicate metrics.
The message in metrics is only useful if it is received. When you are communicating
about metrics, take time to help the listener turn the information into knowledge.
Perhaps ask yourself if you can make a clear visual representation of the information
that highlights the key message of the metric.
We don’t expect that a command of marketing metrics will make your job easier. We
do expect that such knowledge will help you do your job better.
Chapter 13 System of Metrics
415
This page intentionally left blank
BIBLIOGRAPHY
Aaker, David A. (1996). Building Strong Brands, New York: The Free Press.
Aaker, David A. (1991). Managing Brand Equity, New York: The Free Press.
Aaker, David A., and Kevin Lane Keller. (1990). “Consumer Evaluations of Brand
Extensions,” Journal of Marketing, V54 (Jan), 27.
Aaker, David A., and James M. Carman. (1982). “Are You Over Advertising?” Journal
of Advertising Research, 22, 57–70.
Abela, Andrew, Bruce H. Clark, and Tim Ambler. “Marketing Performance
Measurement, Performance, and Learning,” working paper, September 1, 2004.
Abraham, Magid H., and Leonard M. Lodish. (1990). “Getting the Most Out of
Advertising and Promotion,” Harvard Business Review, May–June, 50–58.
Ailawadi, Kusum, Donald Lehmann, and Scott Neslin (2003). “Revenue Premium as an
Outcome Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1-17.
Ailawadi, Kusum, Paul Farris, and Ervin Shames. (1999). “Trade Promotion: Essential
to Selling through Resellers,” Sloan Management Review, Fall.
Ambler, Tim, and Chris Styles. (1995). “Brand Equity: Toward Measures That Matter,”
working paper No. 95-902, London Business School, Centre for Marketing.
Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used
and Where?” Marketing Science Institute, (03-111), working paper, Series issues two
03-002.
Berger, Paul D., Bruce Weinberg, and Richard C. Hanna. (2003). “Customer Lifetime
Value Determination and Strategic Implications for a Cruise-Ship Line,” Database
Marketing and Customer Strategy Management, 11(1), 40-52.
Blattberg, Robert C., and Stephen J. Hoch. (1990). “Database Models and Managerial
Intuition: 50% Model + 50% Manager,” Management Science, 36, No. 8, 887–899.
Borden, Neil H. (1964). “The Concept of the Marketing Mix,” Journal of Advertising
Research, 4, June: 2-7.
Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, Eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” in Does Marketing Measure
Up? Performance Metrics: Practices and Impact, Marketing Science Conference
Summary, No. 05-301.
417
Christen, Markus, Sachin Gupta, John C. Porter, Richard Staelin, and Dick R. Wittink.
(1994). “Using Market-Level Data to Understand Promotion Effects in a Nonlinear
Model,” Journal of Marketing Research, August, Vol. 34, No. 3, 322–334.
Clark, Bruce H., Andrew V. Abela, and Tim Ambler. “Return on Measurement:
Relating Marketing Metrics Practices to Strategic Performance,” working paper,
January 12, 2004.
Dekimpe, Marnik G., and Dominique M. Hanssens. (1995). “The Persistence of
Marketing Effects on Sales,” Marketing Science, 14, 1–21.
Dolan, Robert J., and Hermann Simon. Power Pricing: How Managing Price Transforms
the Bottom Line, New York: The Free Press.
Farris, Paul W., David Reibstein, and Ervin Shames. (1998). “Advertising Budgeting:
A Report from the Field,” New York: American Association of Advertising Agencies.
Forrester, Jay W. (1961). “Advertising: A Problem in Industrial Dynamics,” Harvard
Business Review, March–April, 110.
Forrester, Jay W. (1965), “Modeling of Market and Company Interactions,” Peter D.
Bennet, ed. Marketing and Economic Development, American Marketing Association,
Fall, 353–364.
Gregg, Eric, Paul W. Farris, and Ervin Shames. (revised, 2004). “Perspective on Brand
Equity,” Darden School Technical Notes, UVA-M-0668.
Greyser, Stephen A. (1980). “Marketing Issues,” Journal of Marketing, 47, Winter,
89-93.
Gupta and Lehman. (2003). “Customers As Assets,” Journal of Interactive Marketing,
17(1), 9–24.
Hauser, John, and Gerald Katz. (1998). “Metrics: You Are What You Measure,”
European Management Journal, Vo. 16, No. 5, 517–528.
Interactive Advertising Bureau. Interactive Audience Measurement and Advertising
Campaign Reporting and Audit Guidelines, September 2004, United States Version
6.0b.
Kaplan, R. S., and V.G. Narayanan. (2001). “Measuring and Managing Customer
Profitability.” Journal of Cost Management, September/October: 5–15.
Keiningham, Timothy, Bruce Cooil, Tor Wallin Andreassen, and Lerzan Aksoy (2007).
“A Longitudinal Examination of Net Promoter and Firm Revenue Growth.” Journal of
Marketing, Volume 71, July.
Little, John D.C. (1970). “Models and Managers: The Concept of a Decision Calculus,”
Management Science, 16, No. 8, b-466—b-484.
418
MARKETING METRICS
Lodish, Leonard M. (1997). “J.P. Jones and M.H. Blair on Measuring Advertising Effects
“Another Point of View,” Journal of Advertising Research, September–October, 75–79.
Marn, Michael V., Eric V. Roegner, and Craig C. Zawada. The price advantage. Vol.
286. John Wiley & Sons, 2004.
McGovern, Gail J., David Court, John A. Quelch, and Blair Crawford. (2004). “Bringing
Customers into the Boardroom,” Harvard Business Review, November, 70–80.
Meyer, Christopher (1994), How the Right Measures Help Teams Excel, Harvard
Business Review, May–June, pp. 95–103.
Much, James G., Lee S. Sproull, and Michal Tamuz. (1989). “Learning from Samples of
One or Fewer,” Organizational Science, Vol. 2, No. 1, February, 1–12.
Murphy, Allan H., and Barbara G. Brown. (1984). “A Comparative Evaluation
of Objective and Subjective Weather Forecasts in the United States,” Journal of
Forecasting, Vol. 3, 369–393.
Peppers, D., and M. Rogers. (1997). Enterprise One to One: Tools for Competing in the
Interactive Age, New York: Currency Doubleday.
Pfeifer, P.E., Mark E. Haskins, and Robert M. Conroy. (2005). “Customer Lifetime
Value, Customer Profitability, and the Treatment of Acquisition Spending,” Journal of
Managerial Issues, 11-25.
Poundstone, William. (1993). Prisoner’s Dilemma, New York: Doubleday, 118.
Rangan, V. Kasturi, and Marie Bell. Nestle refrigerated foods: contadina pasta & pizza
(A). Harvard Business School, 1994.
Reichheld, Frederick F., and Earl W. Sasser, Jr. (1990). “Zero Defections: Quality Comes
to Services,” Harvard Business Review, September–October, 105–111.
Reichheld, Frederick F. (2006). The Ultimate Question: Driving Good Profits and True
Growth. Boston: Harvard Business School Publishing Corporation.
Sheth, Jagdish N., and Rajendra S. Sisodia. (2002). “Marketing Productivity Issues and
Analysis,” Journal of Business Research, 55, 349–362.
Tellis, Gerald J., and Doyle L. Weiss. (1995). “Does TV Advertising Really Affect Sales?
The Role of Measures, Models, and Data Aggregation,” Journal of Marketing Research,
Fall, Vol. 24-3.
Wilner, Jack D. (1998). 7 Secrets to Successful Sales Management, Boca Raton, Florida:
CRC Press LLC; 35–36, 42.
Bibliography
419
Zellner, A., H. Kuezenkamp, and M. McAleer. “Simplicity, Inference and Modeling
(Keeping It Sophisticatedly Simple),” Cambridge: Cambridge University Press. (2001).
Zoltners, Andris A., and Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
Complete Guide to Accelerating Sales Force Performance, New York: AMACON.
420
BIBLIOGRAPHY
ENDNOTES
Chapter 1
1. Word Reference, www.wordreference.com. Accessed 22 April 2005.
2. Bartlett, John. (1992). Bartlett’s Familiar Quotations, 16th edition; Justin Kaplan,
general editor.
3. Hauser, John, and Gerald Katz. “Metrics: You are What You Measure,” European
Management Journal, Volume 16 No 5 October 1998.
4. Kaplan, Robert S., and David P. Norton. (1996). Balanced Scorecard, Boston, MA:
Harvard Business School Press.
5. Brady, Diane, with David Kiley and Bureau Reports. “Making Marketing Measure
Up,” Business Week, December 13, 2004, 112–113.
6. Strictly speaking, all the numbers contain some error. For example, share might be
estimated from retail sales to consumers or from shipments to retailers.
7. Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used and
Where?” Marketing Science Institute (03-111), working paper, Series issues two 03-002.
8. Ambler, Tim, Flora Kokkinaki, and Stefano Puntoni. (2004). “Assessing Marketing
Performance: Reasons for Metrics Selection,” Journal of Marketing Management, 20,
475–498.
Chapter 2
1. “Wal-Mart Shopper Update,” Retail Forward, February 2005.
2. “Running Out of Gas,” Business Week, March 28th, 2005.
3. American Marketing Association definition. Accessed 06/08/2005. http://www
.marketingpower.com/live/mg-dictionary.php?SearchFor=market+concentration&
Searched=1.
4. Check the Marketing Evaluations, Inc., Web site for more detail: http://www.qscores
.com/. Accessed 03/03/05.
5. Claritas provides the Prizm analysis. For more details, visit the company Web site:
http://www.clusterbigip1.claritas.com/claritas/Default.jsp. Accessed 03/03/05.
421
6. Reichheld, Fred, The Ultimate Question: Driving Good Profits and True Growth
(Boston: Harvard Business School Publishing Corporation, 2006.)
7. http://www.theultimatequestion.com/theultimatequestion/measuring_netpromoter.asp?groupCode=2
8. Timothy Keiningham, Bruce Cooil, Tor Wallin Andreassen and Lerzan Aksoy
(2007) “A Longitudinal Examination of Net Promoter and Firm Revenue Growth.”
Journal of Marketing, Volume 71, July 2007.
Chapter 3
1. “Running Out of Gas,” Business Week, March 28th, 2005.
2. This formula should be familiar if we consider that the supplier selling price is
merely the cost to that layer of the chain. So this becomes Selling Price = Cost/(1 –
Margin %). This is the same as Sale $ = Cost $ + Margin $.
3. Those familiar with basic economics use the term “marginal cost” to refer to the cost
of an additional unit of output. In this linear cost model, marginal cost is the same
for all units and is equal to the variable cost per unit.
4. Both contribution per unit ($) and contribution margin (%) are closely related to unit
margin ($) and margin (%). The difference is that contribution margins (whether unitor percentage-based) result from a more careful separation of fixed and variable costs.
Chapter 4
1. Rangan, V. Kasturi, and Marie Bell. Nestle Refrigerated Foods: Contadina Pasta &
Pizza (A). Harvard Business School, 1994.
2. Kusum Ailawadi, Donald Lehmann, and Scott Neslin (2003), “Revenue Premium as
an Outcome Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1-17.
3. Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” Does Marketing Measure
Up? Performance Metrics: Practices and Impact, Marketing Science Conference
Summary, No. 05-301.
4. Young and Rubicam can be found at: at: yr.com/BAV. Accessed 7/28/2015.
5. Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” Does Marketing Measure
Up? Performance Metrics: Practices and Impact, Marketing Science Conference
Summary, No. 05-301.
422
MARKETING METRICS
6. See Darden technical note and original research.
7. The information from Bill Moran comes from personal communications with the
authors.
8. Interbrand can be contacted at: http://www.interbrand.com/. Accessed 7/28/2015.
Chapter 5
1. “Vodafone Australia Gains Customers,” Sydney Morning Herald, January 26, 2005.
2. “Atlanta Braves Home Attendance.” Wikipedia, the free encyclopedia. http://en.
wikipedia.org/wiki/Major_League_Baseball_attendance_records
3. Thanks to Gerry Allan, President, Anametrica, Inc. (developer of Web-based tools
for managers) for his work on this section.
4. Pfeifer, P.E., Haskins, M.E., and Conroy, R.M. (2005). “Customer Lifetime Value,
Customer Profitability, and the Treatment of Acquisition Spending,” Journal of
Managerial Issues, 25 pages.
5. Kaplan, R.S., and V.G. Narayanan. (2001). “Measuring and Managing Customer
Profitability,” Journal of Cost Management, September/October, 5–15.
6. Peppers, D., and M. Rogers. (1997). Enterprise One to One: Tools for Competing in
the Interactive Age, New York: Currency Doubleday.
7. Berger, P.D., B. Weinberg, and R. Hanna. (2003). “Customer Lifetime Value
Determination and Strategic Implications for a Cruise-Ship Line,” Database
Marketing and Customer Strategy Management, 11(1), 40–52.
8. Gupta and Lehman. (2003). “Customers as Assets,” Journal of Interactive Marketing,
17(1), 9–24.
Chapter 6
1. Material in Sections 7.1–7.5 is based on a Note on Sales Force Metrics, written by
Eric Larson, Darden MBA 2005.
2. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The Complete Guide to Accelerating Sales Force Performance, New York: AMACON.
3. Wilner, Jack D. (1998). 7 Secrets to Successful Sales Management, Boca Raton,
Florida: CRC Press LLC; 35–36, 42.
Endnotes
423
4. For more on these total allocations, see Zoltners, Andris A., Prabhakant Sinha,
and Greggor A. Zoltners. (2001). The Complete Guide to Accelerating Sales Force
Performance, New York: AMACON.
5. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
Complete Guide to Accelerating Sales Force Performance, New York: AMACON.
6. Dolan, Robert J., and Benson P. Shapiro. “Milford Industries (A),” Harvard
Business School, Case 584-012.
7. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
Complete Guide to Accelerating Sales Force Performance, New York: AMACON.
8. Jones, Eli, Carl Stevens, and Larry Chonko. (2005). Selling ASAP: Art, Science,
Agility, Performance, Mason, Ohio: South Western, 176.
9. Product category volume is also known as weighted distribution.
10. The authors use the term product category volume (PCV) for this metric. However,
this term is not as widely used in industry as all commodity volume (ACV).
Chapter 7
1. Dolan, Robert J., and Hermann Simon. Power Pricing: How Managing Price
Transforms the Bottom Line, New York: The Free Press, 4.
2. Barwise, Patrick, and John U. Farley, “Which Marketing Metrics Are Used and
Where?” Marketing Science Institute, (03-111) 2003, working paper, Series issues
two 03-002.
3. Constant elasticity functions are also called log linear because they can be expressed
as: log Q = log A + elasticity x log (p).
4. In graphing such relationships, economists often plot price on the vertical axis and
quantity demanded on the horizontal axis. When reviewing a graph, managers are
advised to always check the axis definitions.
5. If price elasticity is expressed in shorthand as a positive number, then we do not
need the negative sign in the formula that follows.
6. Poundstone, William. (1993). Prisoner’s Dilemma, New York: Doubleday, 118.
Chapter 8
1. In this context, we use the term “permanent” with some flexibility, recognizing that
even long-term arrangements must be subject to change in response to market and
industry dynamics.
424
MARKETING METRICS
2. Often, contribution can be used as a proxy for profits.
3. Distribution for coupons is used in the sense of postage and insertion costs, rather
than retail and inventory logistics.
4. For a richer discussion, see Ailawadi, Farris, and Shames, Sloan Management
Review, Fall 1999.
5. Roegner, E., M. Marn, and C. Zawada. (2005). “Pricing,” Marketing Management,
Jan/Feb, Vol. 14 (1).
6. “How to Fix Your Pricing if it is Broken,” by Ron Farmer, CEO, Revenue
Technologies for The Professional Pricing Society: http://www.pricingsociety.com/
htmljournal/ 4thquarter2003/article1.htm. Accessed 03/03/05.
7. The following are the two main types of injury contemplated by the Act: (a): Price
discrimination might be used as a predatory pricing tactic, setting prices below cost
to certain customers to harm competition at the supplier’s level. Anti-trust authorities use the same standards applied to predatory pricing claims under the Sherman
Act and the FTC Act to evaluate allegations of price discrimination used for this
purpose. (b) Secondary Line competitive injury: A seller charging competing buyers different prices for the same “commodity” or discriminating in the provision of
“allowances” such as compensation for advertising and other services may be violating the Robinson-Patman Act. This kind of price discrimination can hurt competition by giving favored customers an edge in the market that has nothing to do with
their superior efficiency. However, in the U.S., price discrimination is generally
lawful, particularly if it reflects the different costs of dealing with diverse buyers or
results from a seller’s attempts to meet a competitor’s prices or services. Clearly this
is not intended to be a legal opinion, and legal advice should be sought for a company’s individual circumstances.
Chapter 9
1. Farris, Paul W. (2003). “Getting the Biggest Bang for Your Marketing Buck,”
Measuring and Allocating Marcom Budgets: Seven Expert Points of View, Marketing
Science Institute Monograph.
Chapter 10
1. Known as client-side tagging, beacon, and 1 X 1 clear pixel technology.
2. The interactive Advertising Bureau gives the following definition of ad impression: “A measurement of response from an ad delivery system to an ad request
from the user’s browser, which is filtered from robotic activity and is recorded
Endnotes
425
at a point as late as possible in the process of delivery of the creative material to
the user’s browser—therefore closest to actual opportunity to see by the user.”
Interactive Audience Measurement and Advertising Campaign Reporting and Audit
Guidelines. September 2004, United States Version 6.0b.
3. Emarketer accessed Jan 19th 2015, http://www.emarketer.com/Corporate/Coverage#/
results/1236 and http://www.emarketer.com/Corporate/Coverage#/results/1237.
4. http://www.nielsen.com/us/en/about-us/nielsen-families.html. Accessed 2/1/2015.
5. https://support.google.com/analytics/answer/1009409?hl=en. Accessed 2/1/2015.
Chapter 11
1. Economic value added is a trademark of Stern Stewart Consultants.
2. The weighted average cost of capital, a.k.a. the WACC, is just the percentage return
expected to capital sources. This finance concept is better left to specialist texts, but
to give a simple example, if a third of a firm’s capital comes from the bank at 6%
and two-thirds from shareholders who expect a 9% return, then the WACC is the
weighted average 8%. The WACC will be different for different companies, depending on their structure and risks.
3. Excel has a function to do this quickly, which we explain at the end of the section.
However, it is important to understand what the calculation is doing.
4. A terminal value in a simple calculation might be assumed to be zero or some simple
figure for the sale of the enterprise. More complex calculations consider estimating
future cashflows; where this is done, ask about assumptions and importance. If the
estimated terminal value is a significant area of the analysis, why have you curtailed
the full analyses at this point?
5. Hawkins, Del I., Roger J. Best, and Charles M. Lillis. (1987). “The Nature and
Measurement of Marketing Productivity in Consumer Durables Industries: A Firm
Level Analysis,” Journal of Academy of Marketing Science, Vol. 1, No. 4, 1–8.
Chapter 12
1. Churn = percent of customers lost each year.
2. ACV = all commodity volume, a measure of distribution coverage (refer to
Section 6.6).
426
MARKETING METRICS
Chapter 13
1. An identity is “an equality satisfied by all values of the variables for which the
expression involved in the equality are defined.” American Heritage Dictionary, 2nd
Edition, Houghton Mifflin Company, Boston, 1982.
In finance, economics, and accounting, an identity is “an equality that must be true
regardless of the value of its variables, or a statement that by definition (or construction) must be true.” Where an accounting identity applies, any deviation from the
identity signifies an error in formulation, calculation, or measurement. http://en.
wikipedia.org/wiki/Accounting_identity#cite_note-0
2. Borden, Neil H., Source: “The Concept of the Marketing Mix,” Journal of Advertising
Research, 4, June 1964: 2-7.
3. Zellner, A., H. Kuezenkamp, and M. McAleer. “Simplicity, Inference and Modeling
(Keeping It Sophisticatedly Simple),” Cambridge: Cambridge University Press.
(2001).
4. Ambler, Tim. (2000). Marketing and the Bottom Line: The New Metrics of Corporate
Wealth, London: Prentice Hall.
5. Meyer, Christopher. (1994). “How the Right Measures Help Teams Excel,” Harvard
Business Review.
Endnotes
427
This page intentionally left blank
INDEX
A
AAU (awareness, attitudes, and usage), 43-47
limitations of, 47
purpose of, 43
abandoning customers, 170
abandonment rate, 347-351
A/B testing, 396
accepters, 35
A.C. Nielsen, 212
active users, 357
ACV (all commodity volume), 208-214
ad awareness, 45
ad rank, 343
advertising metrics, 295
AED, 319-323
CPM, 304-305
CPP (cost per point), 306
Dorfman-Steiner Theorem, 322
effective frequency, 316-317
effective reach, 316-317
exposures, 300-303
frequency, 306-309
frequency response functions, 311-314
impressions, 300-303
OTS, 300-303
reach, 306-309
share of voice, 318
AED (advertising elasticity of demand),
319-323
calculating, 320
Dorfman-Steiner Theorem, 322
AMA (American Marketing Association), 5
ANA (Association of National Advertisers), 5
applications, downloads, 355-356
applying neuroscience to marketing, 59
ARPU (average revenue per use), 356
ASN (Augmented Social Network), 354
average acquisition cost, 180-181
average frequency, 308
average margin, 84-86
average price per unit, 87-92
calculating, 89
limitations of, 92
price per statistical unit, 90-92
purpose of, 88
SKUs, 88
average retention cost, 180-181
awareness, 45
B
balancing workload, 192
baseline sales, 272, 275-282
BDI (brand development index), 30-31
calculating, 31
purpose of, 30-31
bounce rate, 349-350
Brand Asset Valuator tool, 141-142
brand equity index, 140
brand equity metrics, 137-146
Brand Asset Valuator tool, 141-142
brand equity index, 140
Brand Equity Ten, 139
429
brand identity, 146
brand valuation, 145-147
conjoint analysis, 146
purpose of, 137-138
royalty rates, 143
Brand Equity Ten, 139
brand penetration, calculating, 33
brand valuation, 145-147
BrandZ, 143
breakage, 285
break-even analysis, 103-107
break-even point, calculating, 104-107
limitations of, 107
payback period, 108
purpose of, 104
break-even point, calculating, 104-107
breaking down sales, 404
buying power, 192
C
CAGR (compound annual growth rate),
calculating, 127, 131
calculating
abandonment rate, 347-351
active users, 357
ACV, 208-214
AED, 320
average acquisition cost, 181
average price per unit, 89
average retention cost, 181
baseline sales, 276-281
BDI, 31
bounce rate, 350
brand penetration, 33
CAGR, 127, 131
cannibalization rate, 132
channel margins, 79-81
clickthrough rate, 336
CLV, 173
conversion rate, 348
430
MARKETING METRICS
cost per click, 341
cost per impression, 341
cost per order, 341
coupon redemption rate, 284
CP, 166-168
CPM, 305
downloads, 355
EBITDA, 367
economic profit, 370-371
effectiveness of sales force, 196-197
effective reach, 316
frequency response functions, 312-314
future value, 130
hits, 331
impressions, 300
margin on sales, 71-73
market penetration, 33
MROI, 377
net profit, 366
NPV, 376
numeric distribution, 208
optimal price, 246-253
pageviews, 331
pass-through, 287-288
PCV, 210
percentage margins, 72
percent good value, 232-233
PLV, 177
price elasticity, 239-242
price premium, 228-229
price waterfall of a product, 290
relative market share, 27
repeat volume, 119
reservation price, 232-233
residual price elasticity, 260-262
return on sales, 367
revenue market share, 24
rich media display time, 333-334
rich media interaction time, 333-334
ROI, 368
sales force compensation, 200
sales goals, 194
selling costs, 100-101
session length, 357
share of requirements, 36
share of voice, 318
sole usage percentage, 38
target revenue, 109
target volume, 109
total cost, 93-96
total distribution, 211
total sales, 276-279
unit margin, 72
unit market share, 23
usage index, 39-40
visits, 347
workload, 191-192
year-on-year growth, 127
cannibalization
fair share draw, 135
limitations of, 136-140
weighted contribution margin, 134
cannibalization rate, 132-136
carryover effect, 381
category performance ratio, 211
CDI (category development index), 31-32
channel margins, 77-86
average margin, 84-86
calculating, 79-81
hybrid channel margins, 83
limitations of, 83
purpose of, 78
clickstreams, 346
clickthrough rate, 336-338
cost per click, 339-343
maximum cost per click, 343
cluster analysis, 153
CLV (customer lifetime value), 9, 171-175
calculating, 173
cohort and incubate approach, 172-176
infinite horizon assumption, 176
present value, 172
versus PLV, 177-179
with initial margin, 175
cohort and incubate approach, 172-176
cold leads, 204
Common Language Project, 5
comparing markup and margin, 75-77
compensation, 199-201
compensatory decision processes, 149
compounding, 130
concepts, 4
conjoint analysis, 146-151
decision-making processes, 149-151
limitations of, 151-153
purpose of, 148
segmentation
cluster analysis, 153
limitations of, 154-156
volume projection, 155
constant elasticity, 242-244
constructs, 4
consumer off-take, 219
contribution analysis, 103-107
conversion rate, 348
cookies, 349
copy tests, 397
cost per click, 339-343
cost per impression, 339-343
cost per order, 339-343
counting customers, 160-164
coupon redemption rate, 283-284
CP (customer profit)
calculating, 166-168
limitations of, 169-171
purpose of, 165
CPE (cost per engagement), 359
CPM (cost per thousand impressions),
304-305
CPP (cost per point), 306
“cross” price elasticity, 258-260
Index
431
customers
abandoning, 170
average acquisition cost, 180-181
average retention cost, 180-181
CLV, 171-175
calculating, 173
cohort and incubate approach, 172-176
infinite horizon assumption, 176
present value, 172
with initial margin, 175
counting, 160-161, 163-164
CP, 165-170
defining, 163
PLV
calculating, 177
limitations of, 178-179
PLV (prospect lifetime value), 177-179
recency, 162
relationships, 164
retention rate, 162-163
segmentation, 152-154
tiers, 166
under-servicing, 191
customer satisfaction, 48-51
limitations, 51
purpose of, 49
D
dashboards, 391-393
data, 5, 43
decision-making processes, 149-151
decomposition of market share, 34
deductions, 218, 292
defining customers, 163
discounted trial, 126
discounts, 291
distribution metrics, 206-212
ACV, 208-214
numeric distribution, 207-208
PCV (product category volume), 210
total distribution, 211
432
MARKETING METRICS
districts, 194
diverted merchandise, 219
Dorfman-Steiner Theorem, 322
double jeopardy, 37, 40
downloads, 355-356
DPP (direct product profitability), 221-222
Drucker, Peter, 67
DuPont Model, 401-402
durability, 140
E
EBITDA (earnings before interest, taxes,
depreciation, and amortization), 367
economic profit, 369-370
EDA (electrodermal activity), 60
EDLP (everyday low prices), 292
EEG (electroencephalography), 57
effective frequency, 316-317
effective market share, 140
effectiveness of sales force, 196-198
effective reach, 316-317
elasticity, 226, 319-323
elimination-by-aspect decision making, 150
email metrics, 358-360. See also mobile
metrics
EMG (electromyography), 60
engagement, 359-361
estimating MROI, 380
EVA (economic value added), 9, 369-370
evaluating multi-period investments, 371-375
internal rate of return, 374
NPV, 373
payback, 372
ever-tried, 126
evoked set, 127
examples of neuro-marketing, 61
expected lift, 400
exposures, 300-303
frequency response functions, 311-314
external overlap, 310
eye tracking, 59-60
F
H
facing, 212
FACS (Facial Action Coding System), 58-59
fair share draw, 135
FIFO (first in, first out), 217
financial metrics, 393
fixed costs, 93-97, 102
fMRI (functional magnetic resonance
imaging), 58
forced trial, 126
forecasting
sales, 194
trial volume, 118
formulas, unit market share metric, 23
frequency, 306-309
effective frequency, 316-317
frequency response functions, 311-314
friends, 352-353
future value, calculating, 130
Herfindahl index, 29
Hierarchy of Effects, 43-47
HI-LO pricing, 292
hits, 329-331
hybrid channel margins, 83-86
G
GMROII (gross margin return on inventory
investment), 219
goodwill, 138-139
Google Analytics, 351
gross margin, 77
Gross model, 396-400
growth, 127-130
CAGR, calculating, 131
compounding, 130
future value, calculating, 130
percentage growth, 128
same stores growth, 128-131
GRPs (gross rating points), 301
GSR (Galvanic Skin Response), 60
I
IAT (Implicit Association Test), 60
identities, reasons for using, 413
diagnostic purposes, 405
eliminating error, 406
estimating metrics, 408
impressions, 296, 300-303, 329-331
calculating, 300
clickthrough rate, 336-338
cost per impression, 339-343
CPM, 304-305
increasing value of information, 6
incremental sales, 272
calculating, 276
infinite horizon assumption, 176
information, 5, 393-394
intentions, 46
interactions, 335
internal rate of return, 374
inventory days, 216
inventory turns, 215
invoice price, 289
J-K-L
Kaplan, Robert, 167
learning curve response, 312-315
LIFO (last in, first out), 217
likeability, 48
Index
433
likes, 352-353
limitations
of AAU, 47
of average price per unit, 92
of break-even analysis, 107
of cannibalization metric, 136-140
of channel margins, 83
of conjoint analysis, 151-153
of CP metric, 169-171
of customer satisfaction metric, 51
of margin of sales, 74-76
of net promoter metric, 53
of neuro-marketing, 62
of penetration metric, 34
of PLV, 178-179
of price elasticity, 245
of price premium, 231
of revenue market share, 24
of segmentation, 154-156
of selling costs, 102
of share of requirements metric, 37
of supply chain metrics, 217-222
of total cost, 97
of unit market share, 24
of willingness to recommend, 51
of willingness to search metric, 56
linear demand, 234-237
list price, 289
logistics tracking, 213
long-term effects of promotion, 282
Lord Kelvin, 2
loss aversion, 394
M
mail-in rebates, 285-289
margin of sales, 413
calculating, 71-73
gross margin, 77
434
MARKETING METRICS
limitations of, 74-76
purpose of, 71
versus markup, 75
markdown, 219
market concentration, 26-29
purpose of, 26
marketing metrics x-ray, 383-392
marketing mix models, 408-412
marketing spending, 99-102
calculating, 100-101
purpose of, 99
market penetration, calculating, 33
market share metrics, 18-21
BDI, 30-31
market concentration, 26-29
penetration, 32-34
relative market share, 26-29
revenue market share
calculating, 24
purpose of, 23
share of requirements, 35-38
calculating, 36
limitations of, 37
unit market share, 23-24
usage index, 39
market share rank, 30
markup versus margin, 75-77
“Marlboro Friday”, 391
MASB (Marketing Accountability Standards
Board), 5
mastering metrics, 7
maximum cost per click, 343
measurements, validating, 7
MEROMI (media exposure return on
marketing investment), 381
metrics
defined, 1
mastering, 7
“top ten”, 8
Microsoft Excel, NPV calculator, 376
misshipments, 218
mobile metrics, 356-357
modeling firm performance
DuPont Model, 401-402
sales model, 403
MROI (marketing return on investment), 364,
376-380
MRP (maximum reservation price), 235
MSI (Marketing Science Institute), 5
multi-period investments, 371-375
internal rate of return, 374
NPV, 373
payback, 372
MWB (maximum willingness to buy), 236
N
net out-of-stocks, 215
net price, 289
net profit, 366-367
net promoter, 52-53
net promoter metric, limitations of, 53
net reach, 306-309
neuro-marketing, 9
example of, 61
limitations of, 62
neuroscience measures, 57
applying to marketing problems, 59
EEG, 57
eye tracking, 59-60
FACS (Facial Action Coding System), 58-59
fMRI, 58
purpose of, 60
non-compensatory decision processes, 149
NOPAT (net operating profit after tax), 368
NPV (net present value), 373
numeric distribution, 207-208
O
obsolescence, 218
online metrics, 9
operational definitions, 4
optimal price
calculating, 246-253
linear demand functions, 249-252
price discrimination, 254-256
purpose of, 246
relative to gross margin, 253
organization of this book, 10-15
OTS (opportunities to see), 300-303
outcomes per friend, 353
out-of-stocks, 214
“own” price elasticity, 258-260
P
PageRank, 344
pageviews, 329-331
paid search marketing, 342
pass-through, 286-287
payback, 372
payback period, 108
PCV (product category volume), 210
penetration, 32-34
perceived quality/esteem, 46
perceived value for money, 46
percentage growth, 127-130
percentage margins, calculating, 72
percent good value, 232-237
pipeline analysis, 202-205
PLV (prospect lifetime value), 177-179
calculating, 177
limitations of, 178-179
present value, 172
price discrimination, 254-256, 292-294
Index
435
price elasticity, 238-244
calculating, 239-242
constant elasticity, 242-244
limitations of, 245
residual price elasticity, 258-267
price per statistical unit, 90-92
price premium, 228-231
calculating, 228-229
limitations of, 231
purpose of, 228
theoretical price premium, 232
price waterfall, 288-292
calculating, 290
purpose of, 289
pricing strategy
optimal price, 245-256
calculating, 246-253
linear demand functions, 249-252
price discrimination, 254-256
purpose of, 246
relative to gross margin, 253
percent good value, 232-237
price elasticity, 238-244
calculating, 239-242
constant elasticity, 242-244
limitations of, 245
price premium, 230-231
calculating, 228-229
purpose of, 228
prisoner’s dilemma, 262-267
reservation price, 232-237
calculating, 232-233
linear demand, 234-237
residual price elasticity, 259
calculating, 260
purpose of, 258
theoretical price premium, 232
primary line competitive injury, 257
prisoner’s dilemma, 262-267
436
MARKETING METRICS
product strategies
lifecycle, 131
projected volume, repeat volume, 119
product strategy metrics, 112-113
cannibalization
fair share draw, 135
limitations of, 136-140
weighted contribution margin, 134
cannibalization rate, 132-136
trial volume, 116-118
profitability, 219-222
CP, 165-170
DPP, 221-222
economic profit, 369-370
EVA, 9
GMROII, 219
markdown, 219
of promotions, 279-282
shopping basket margin, 223, 223-224
projected volume
repeat volume, 119
trial volume, 116-118
projection of sales, 116
promotion, 272
baseline sales, 275-282
coupon redemption rate, calculating, 284
long-term effects of, 282
mail-in rebates, 285-289
pass-through, 286-287
price waterfall, 288-292
calculating, 290
purpose of, 289
profitability, 279-282
redemption rate, 283-287
total sales, calculating, 276-279
purchase intentions, 46
purpose
of AAU, 43
of average price per unit, 88
of BDI, 30-31
of brand equity metrics, 137-138
of break-even analysis, 104
of channel margins, 78
of conjoint analysis, 148
of CP metric, 165
of customer satisfaction metric, 49
of margin on sales, 71
of market concentration metric, 26
of marketing spending, 99
of net promoter metric, 52-53
of neuro-marketing, 60
of optimal price, 246
of pipeline analysis, 202
of price premium, 228
of price waterfall, 289
of relative market share metric, 26
of residual price elasticity, 258
of revenue market share, 23
of segmentation, 153
of selling costs, 99
of target revenue, 108
of target volume, 108
of total cost, 93
of unit market share, 23
of willingness to recommend metric, 49
of willingness to search metric, 55
push marketing, 185
Q-R
Q Scores, 48
quality scores, 398
rain checks, 218
rate of return, 363
reach, 306-309
effective reach, 316-317
recency, 162
redemption rate, 283-287
calculating, 284
mail-in rebates, 285-289
relationships, 164
relative market share, 26-29
calculating, 27
purpose of, 26
relative perceived quality, 46
relative price, 140, 228-231
reliability, 6
repeat rate, 38
repeat volume, 119
repurchase rate, 38
resellers, 286-287
reservation price, 232-237
calculating, 232-233
linear demand, 234-237
residual income. See economic profit
residual price elasticity, 258-267
calculating, 260-262
purpose of, 258
response latencies, 60
retention rate, 162-163
return on sales, 366-367
revenue market share, 23-24
calculating, 24
limitations of, 24
purpose of, 23
rich media display time, 333-334
rich media interaction time, 333-334
risk aversion, 394
ROA (return on assets), 368
Robinson-Patman Act, 257
ROI (return on investment), 368-369
MROI, 376-380
ROIC (return on invested capital), 368
ROIMI (return on incremental marketing
investment), 378
RONA (return on net assets), 368
royalty rates, 143
Index
437
S
sales force metrics
compensation, 199-201
effectiveness of personnel, 196-198
pipeline analysis, 202-205
sales goals, 193-194
territories, 190
sales goals, 193-194
calculating, 194
sales potential, 193
same stores growth, 128-131
SCR (Skin Conductance Response), 60
S-curve, 409
search engine optimization, 344
search engines, 342
secondary line competitive injury, 257
segmentation, 152-154
cluster analysis, 153
limitations of, 154-156
purpose of, 153
selling costs, 99-102
calculating, 100-101
limitations of, 102
purpose of, 99
slotting allowances, 103
served markets, 25
service levels, 215
session length, 357
setting sales goals, 193-194
Shames, Erv, 390
share of category, 30
share of requirements, 35-38
calculating, 36
limitations of, 37
share of shelf, 212
share of voice, 318
shopping basket margin, 223
shopping carts, 347
shrinkage, 218
SKUs (stock keeping units), 88
slotting allowances, 103
438
MARKETING METRICS
social media metrics, 352-353
sole usage percentage, 38
supply chain metrics, 213-218
inventory days, 216
inventory turns, 215
limitations of, 217-222
net out-of-stocks, 215
out-of-stocks, 214
service levels, 215
system of identities, reasons for using
diagnostic purposes, 405
eliminating error, 406
estimating metrics, 408
T
target revenue, 108-109
calculating, 109
purpose of, 108
target volume, 108-109
calculating, 109
purpose of, 108
Teixeira, Thales, 58-59
territories, 190
testing, 6, 395-399
A/B testing, 396
copy tests, 397
expected lift, 399-400
Gross model, 396-400
quality scores, 398
theoretical price premium, 232
three firm concentration ratio, 28
threshold response, 312-315
tiers of customers, 166
top of mind, 45
“top ten” metrics, 8
total cost, 93-97
calculating, 93-96
limitations of, 97
purpose of, 93
total distribution, 211
total sales, calculating, 276-279
tracking data, 43
translating concepts into metrics, 4
trial rate, 115
trial volume, 116-118
TRPs (target rating points), 301-304
U
unit margin, calculating, 72
unit market share, 23-24
calculating, 23
limitations, 24
usage, 46
usage index, 41-42
and double jeopardy, 40
calculating, 39-40
willingness to search, 54
limitations of, 56
purpose of, 55
workload, 190
calculating, 191-192
WTP (willingness to pay), 5
X-Y-Z
year-on-year growth, calculating, 127
Young & Rubicam, 141
Zellner, Arnold, 409
V
validating measurements, 7
value of information, 5-7, 393-394
variable costs, 93-97, 102
verifying unit margin, 72
version quality, assessing, 399
video interactions, 335
visitors, 345
visits, 344-348
bounce rate, 349-350
cookies, 348
volume projections, 115
W
warm leads, 204
wear-in, 315
wear-out, 315
weighted contribution margin, 134
weight index, 39-42
“whale curves”, 167
willingness to recommend, 48
limitations of, 51
purpose of, 49-50
Index
439